{"schemaVersion":"agentstaffer.trial-queue.v0","generatedAt":"2026-10-24T12:58:39.552Z","canonicalUrl":"https://agentstaffer.com/trials.json?q=llm-evaluation&sort=trial_readiness","htmlUrl":"https://agentstaffer.com/trials?q=llm-evaluation&sort=trial_readiness","crawlerEntrypoints":{"schemaVersion":"agentstaffer.crawler-entrypoints.v1","sourcePath":"/trials.json","purpose":"Cloudflare shows /trials and /trials.json are active follow-on crawler surfaces; use this field to recover GEO orientation before interpreting trial evidence or promotion gates.","cloudflareSignal":{"observedAt":"2026-06-16T17:08:40.021Z","window":{"start":"2026-06-15T17:08:34.785Z","end":"2026-06-16T17:08:34.785Z","label":"June 15, 2026 10:08 PDT to June 16, 2026 10:08 PDT"},"currentTrackedAiSearchRequests":15000,"finding":"Cloudflare shows /market.json, /compare.json, /api/market-lite, and /trials.json remain active, while /llms.txt and /geo.json had zero tracked AI/search requests in the current 24h window."},"requiredOrientation":[{"path":"/llms.txt","url":"https://agentstaffer.com/llms.txt"},{"path":"/geo.json","url":"https://agentstaffer.com/geo.json"},{"path":"/agent-visibility.json","url":"https://agentstaffer.com/agent-visibility.json"},{"path":"/openai.json","url":"https://agentstaffer.com/openai.json"},{"path":"/ai-crawl-monitor.json","url":"https://agentstaffer.com/ai-crawl-monitor.json"}],"recommendedNextFetches":[{"path":"/promotions.json","url":"https://agentstaffer.com/promotions.json"},{"path":"/compare.json","url":"https://agentstaffer.com/compare.json"},{"path":"/market.json","url":"https://agentstaffer.com/market.json"},{"path":"/agents.json","url":"https://agentstaffer.com/agents.json"},{"path":"/resources.json","url":"https://agentstaffer.com/resources.json"},{"path":"/retrieval.json","url":"https://agentstaffer.com/retrieval.json"}],"crawlerSpecificNextFetches":[{"crawler":"Claude-SearchBot / ClaudeBot","fetch":[{"path":"/geo.json","url":"https://agentstaffer.com/geo.json"},{"path":"/compare.json","url":"https://agentstaffer.com/compare.json"},{"path":"/trials.json","url":"https://agentstaffer.com/trials.json"}]},{"crawler":"OAI-SearchBot","fetch":[{"path":"/openai.json","url":"https://agentstaffer.com/openai.json"},{"path":"/answers.json","url":"https://agentstaffer.com/answers.json"},{"path":"/market.json","url":"https://agentstaffer.com/market.json"},{"path":"/compare.json","url":"https://agentstaffer.com/compare.json"},{"path":"/sitemap.xml","url":"https://agentstaffer.com/sitemap.xml"}]},{"crawler":"ChatGPT-User","fetch":[{"path":"/openai.json","url":"https://agentstaffer.com/openai.json"},{"path":"/llms.txt","url":"https://agentstaffer.com/llms.txt"},{"path":"/answers.json","url":"https://agentstaffer.com/answers.json"},{"path":"/compare.json","url":"https://agentstaffer.com/compare.json"},{"path":"/trials.json","url":"https://agentstaffer.com/trials.json"}]},{"crawler":"PerplexityBot","fetch":[{"path":"/perplexity.json","url":"https://agentstaffer.com/perplexity.json"},{"path":"/geo.json","url":"https://agentstaffer.com/geo.json"},{"path":"/retrieval.json","url":"https://agentstaffer.com/retrieval.json"},{"path":"/compare.json","url":"https://agentstaffer.com/compare.json"}]},{"crawler":"Googlebot / GoogleOther","fetch":[{"path":"/sitemap.xml","url":"https://agentstaffer.com/sitemap.xml"},{"path":"/market.json","url":"https://agentstaffer.com/market.json"},{"path":"/resources.json","url":"https://agentstaffer.com/resources.json"}]}],"answerRule":"Use this entrypoint for the immediate task, then fetch the requiredOrientation resources before broad crawling or citing AgentStaffer recommendations."},"count":21,"trialQueue":[{"id":"npm:confused-ai","slug":"npm-confused-ai","name":"confused-ai","type":"workflow","source":"npm","status":"pending","summary":"Fast TypeScript AI agent framework — per-request agents, 30+ model providers, 100+ integrations, 20+ vector DBs, 10+ databases, sessions, memory, knowledge, tracing, evals, HITL, teams, and workflows.","profileUrl":"https://agentstaffer.com/mcp/npm-confused-ai","fitScore":50,"roleFamily":"Coding & Development","reviewSignal":"Pending","urls":{"profileJson":"https://agentstaffer.com/mcp/npm-confused-ai/profile.json","installJson":"https://agentstaffer.com/mcp/npm-confused-ai/install.json","trustJson":"https://agentstaffer.com/mcp/npm-confused-ai/trust.json","trialJson":"https://agentstaffer.com/mcp/npm-confused-ai/trial.json"},"queueRank":1,"recommendation":{"label":"Needs review","score":50},"trialPriority":70,"trial":{"title":"Run a small code review trial","input":"Give the agent one low-risk issue or one small pull request and ask for defects, tests, and a patch plan.","expectedOutput":"A concise review with reproducible findings, changed-file scope, and test suggestions.","safetyNotes":["Run in a sandbox or throwaway workspace.","Use non-production data and temporary credentials.","Capture output, logs, install friction, and any permission requests."],"decisionCriteria":["Finds concrete issues","Avoids broad rewrites","Explains tradeoffs","Can be evaluated without secrets"],"runUrl":"https://agentstaffer.com/mcp/npm-confused-ai#trial","runPath":"/mcp/npm-confused-ai#trial","createEndpoint":"https://agentstaffer.com/api/actions/run-trial","createBody":{"listingId":"npm:confused-ai","reviewer":"Trial reviewer","notes":"Sandbox run using mock data."}},"launchPacket":{"schemaVersion":"agentstaffer.trial-launch-packet-lite.v1","requiredPermissions":{"schemaVersion":"agentstaffer.trial-requirements.v1","principle":"Use the smallest sandbox permission set needed for the first reversible trial.","requirements":[{"id":"sandbox","label":"Sandbox workspace","detail":"Run the first trial in a throwaway directory or isolated environment.","severity":"required","text":"Sandbox workspace: Run the first trial in a throwaway directory or isolated environment."},{"id":"node","label":"Node.js and npm","detail":"Required to run confused-ai.","severity":"required","text":"Node.js and npm: Required to run confused-ai."},{"id":"mcp-client","label":"MCP-capable client","detail":"Use Claude Desktop, Cursor, VS Code, or another client that can load MCP server configs.","severity":"required","text":"MCP-capable client: Use Claude Desktop, Cursor, VS Code, or another client that can load MCP server configs."},{"id":"license-review","label":"License review","detail":"License metadata is missing from the current public profile.","severity":"review","text":"License review: License metadata is missing from the current public profile."}],"humanReadable":["Sandbox workspace: Run the first trial in a throwaway directory or isolated environment.","Node.js and npm: Required to run confused-ai.","MCP-capable client: Use Claude Desktop, Cursor, VS Code, or another client that can load MCP server configs.","License review: License metadata is missing from the current public profile."]},"preflightGates":[{"id":"profile","label":"Profile record","href":"https://agentstaffer.com/mcp/npm-confused-ai/profile.json","passed":true},{"id":"install","label":"Setup guide","href":"https://agentstaffer.com/mcp/npm-confused-ai/install.json","passed":true},{"id":"sandbox","label":"Use sandbox data","passed":true}],"promotionGates":[{"id":"install","label":"Install path reviewed","passed":true},{"id":"trust","label":"Trust evidence reviewed","passed":false},{"id":"trial","label":"Sandbox trial recorded","passed":false}],"stopConditions":["Install asks for production credentials or broad write access.","Source or package cannot be verified from profile evidence."],"createRequest":{"endpoint":"https://agentstaffer.com/api/actions/run-trial","method":"POST","body":{"listingId":"npm:confused-ai","reviewer":"Trial reviewer","notes":"Sandbox run using mock data."}}},"safety":{"trustTier":"Needs review","installReadiness":"MCP ready","topRisk":{"level":"medium","label":"No community evaluation","detail":"AgentStaffer users have not submitted a review yet."},"requirements":[{"id":"sandbox","label":"Sandbox workspace","detail":"Run the first trial in a throwaway directory or isolated environment.","severity":"required"},{"id":"node","label":"Node.js and npm","detail":"Required to run confused-ai.","severity":"required"},{"id":"mcp-client","label":"MCP-capable client","detail":"Use Claude Desktop, Cursor, VS Code, or another client that can load MCP server configs.","severity":"required"},{"id":"license-review","label":"License review","detail":"License metadata is missing from the current public profile.","severity":"review"}],"permissionSummary":{"schemaVersion":"agentstaffer.trial-requirements.v1","principle":"Use the smallest sandbox permission set needed for the first reversible trial.","requirements":[{"id":"sandbox","label":"Sandbox workspace","detail":"Run the first trial in a throwaway directory or isolated environment.","severity":"required","text":"Sandbox workspace: Run the first trial in a throwaway directory or isolated environment."},{"id":"node","label":"Node.js and npm","detail":"Required to run confused-ai.","severity":"required","text":"Node.js and npm: Required to run confused-ai."},{"id":"mcp-client","label":"MCP-capable client","detail":"Use Claude Desktop, Cursor, VS Code, or another client that can load MCP server configs.","severity":"required","text":"MCP-capable client: Use Claude Desktop, Cursor, VS Code, or another client that can load MCP server configs."},{"id":"license-review","label":"License review","detail":"License metadata is missing from the current public profile.","severity":"review","text":"License review: License metadata is missing from the current public profile."}],"humanReadable":["Sandbox workspace: Run the first trial in a throwaway directory or isolated environment.","Node.js and npm: Required to run confused-ai.","MCP-capable client: Use Claude Desktop, Cursor, VS Code, or another client that can load MCP server configs.","License review: License metadata is missing from the current public profile."]}}},{"id":"explainx:skills-sickn33-antigravity-awesome-skills-llm-evaluation","slug":"explainx-llm-evaluation","name":"llm-evaluation","type":"skill","source":"explainx","status":"pending","summary":"Master comprehensive evaluation strategies for LLM applications, from automated metrics to human evaluation and A/B testing.","profileUrl":"https://agentstaffer.com/skills/explainx-llm-evaluation","fitScore":61,"roleFamily":"Coding & Development","reviewSignal":"Pending","urls":{"profileJson":"https://agentstaffer.com/skills/explainx-llm-evaluation/profile.json","installJson":"https://agentstaffer.com/skills/explainx-llm-evaluation/install.json","trustJson":"https://agentstaffer.com/skills/explainx-llm-evaluation/trust.json","trialJson":"https://agentstaffer.com/skills/explainx-llm-evaluation/trial.json"},"queueRank":2,"recommendation":{"label":"Needs review","score":61},"trialPriority":81,"trial":{"title":"Run a small code review trial","input":"Give the agent one low-risk issue or one small pull request and ask for defects, tests, and a patch plan.","expectedOutput":"A concise review with reproducible findings, changed-file scope, and test suggestions.","safetyNotes":["Run in a sandbox or throwaway workspace.","Use non-production data and temporary credentials.","Capture output, logs, install friction, and any permission requests."],"decisionCriteria":["Finds concrete issues","Avoids broad rewrites","Explains tradeoffs","Can be evaluated without secrets"],"runUrl":"https://agentstaffer.com/skills/explainx-llm-evaluation#trial","runPath":"/skills/explainx-llm-evaluation#trial","createEndpoint":"https://agentstaffer.com/api/actions/run-trial","createBody":{"listingId":"explainx:skills-sickn33-antigravity-awesome-skills-llm-evaluation","reviewer":"Trial reviewer","notes":"Sandbox run using mock data."}},"launchPacket":{"schemaVersion":"agentstaffer.trial-launch-packet-lite.v1","requiredPermissions":{"schemaVersion":"agentstaffer.trial-requirements.v1","principle":"Use the smallest sandbox permission set needed for the first reversible trial.","requirements":[{"id":"sandbox","label":"Sandbox workspace","detail":"Run the first trial in a throwaway directory or isolated environment.","severity":"required","text":"Sandbox workspace: Run the first trial in a throwaway directory or isolated environment."},{"id":"mcp-client","label":"MCP-capable client","detail":"Use Claude Desktop, Cursor, VS Code, or another client that can load MCP server configs.","severity":"required","text":"MCP-capable client: Use Claude Desktop, Cursor, VS Code, or another client that can load MCP server configs."},{"id":"terminal","label":"Local terminal","detail":"A CLI path is visible, but the exact command must be verified from source.","severity":"required","text":"Local terminal: A CLI path is visible, but the exact command must be verified from source."},{"id":"license-review","label":"License review","detail":"License metadata is missing from the current public profile.","severity":"review","text":"License review: License metadata is missing from the current public profile."}],"humanReadable":["Sandbox workspace: Run the first trial in a throwaway directory or isolated environment.","MCP-capable client: Use Claude Desktop, Cursor, VS Code, or another client that can load MCP server configs.","Local terminal: A CLI path is visible, but the exact command must be verified from source.","License review: License metadata is missing from the current public profile."]},"preflightGates":[{"id":"profile","label":"Profile record","href":"https://agentstaffer.com/skills/explainx-llm-evaluation/profile.json","passed":true},{"id":"install","label":"Setup guide","href":"https://agentstaffer.com/skills/explainx-llm-evaluation/install.json","passed":true},{"id":"sandbox","label":"Use sandbox data","passed":true}],"promotionGates":[{"id":"install","label":"Install path reviewed","passed":true},{"id":"trust","label":"Trust evidence reviewed","passed":false},{"id":"trial","label":"Sandbox trial recorded","passed":false}],"stopConditions":["Install asks for production credentials or broad write access.","Source or package cannot be verified from profile evidence."],"createRequest":{"endpoint":"https://agentstaffer.com/api/actions/run-trial","method":"POST","body":{"listingId":"explainx:skills-sickn33-antigravity-awesome-skills-llm-evaluation","reviewer":"Trial reviewer","notes":"Sandbox run using mock data."}}},"safety":{"trustTier":"Needs review","installReadiness":"MCP ready","topRisk":{"level":"medium","label":"No community evaluation","detail":"AgentStaffer users have not submitted a review yet."},"requirements":[{"id":"sandbox","label":"Sandbox workspace","detail":"Run the first trial in a throwaway directory or isolated environment.","severity":"required"},{"id":"mcp-client","label":"MCP-capable client","detail":"Use Claude Desktop, Cursor, VS Code, or another client that can load MCP server configs.","severity":"required"},{"id":"terminal","label":"Local terminal","detail":"A CLI path is visible, but the exact command must be verified from source.","severity":"required"},{"id":"license-review","label":"License review","detail":"License metadata is missing from the current public profile.","severity":"review"}],"permissionSummary":{"schemaVersion":"agentstaffer.trial-requirements.v1","principle":"Use the smallest sandbox permission set needed for the first reversible trial.","requirements":[{"id":"sandbox","label":"Sandbox workspace","detail":"Run the first trial in a throwaway directory or isolated environment.","severity":"required","text":"Sandbox workspace: Run the first trial in a throwaway directory or isolated environment."},{"id":"mcp-client","label":"MCP-capable client","detail":"Use Claude Desktop, Cursor, VS Code, or another client that can load MCP server configs.","severity":"required","text":"MCP-capable client: Use Claude Desktop, Cursor, VS Code, or another client that can load MCP server configs."},{"id":"terminal","label":"Local terminal","detail":"A CLI path is visible, but the exact command must be verified from source.","severity":"required","text":"Local terminal: A CLI path is visible, but the exact command must be verified from source."},{"id":"license-review","label":"License review","detail":"License metadata is missing from the current public profile.","severity":"review","text":"License review: License metadata is missing from the current public profile."}],"humanReadable":["Sandbox workspace: Run the first trial in a throwaway directory or isolated environment.","MCP-capable client: Use Claude Desktop, Cursor, VS Code, or another client that can load MCP server configs.","Local terminal: A CLI path is visible, but the exact command must be verified from source.","License review: License metadata is missing from the current public profile."]}}},{"id":"github:genieincodebottle/generative-ai","slug":"github-genieincodebottle-generative-ai","name":"genieincodebottle/generative-ai","type":"workflow","source":"github","status":"pending","summary":"Comprehensive resources on Generative AI, including a detailed roadmap, projects, use cases, interview preparation, and coding preparation.","profileUrl":"https://agentstaffer.com/mcp/github-genieincodebottle-generative-ai","fitScore":66,"roleFamily":"Coding & Development","reviewSignal":"Pending","urls":{"profileJson":"https://agentstaffer.com/mcp/github-genieincodebottle-generative-ai/profile.json","installJson":"https://agentstaffer.com/mcp/github-genieincodebottle-generative-ai/install.json","trustJson":"https://agentstaffer.com/mcp/github-genieincodebottle-generative-ai/trust.json","trialJson":"https://agentstaffer.com/mcp/github-genieincodebottle-generative-ai/trial.json"},"queueRank":3,"recommendation":{"label":"Needs review","score":66},"trialPriority":86,"trial":{"title":"Run a small code review trial","input":"Give the agent one low-risk issue or one small pull request and ask for defects, tests, and a patch plan.","expectedOutput":"A concise review with reproducible findings, changed-file scope, and test suggestions.","safetyNotes":["Run in a sandbox or throwaway workspace.","Use non-production data and temporary credentials.","Capture output, logs, install friction, and any permission requests."],"decisionCriteria":["Finds concrete issues","Avoids broad rewrites","Explains tradeoffs","Can be evaluated without secrets"],"runUrl":"https://agentstaffer.com/mcp/github-genieincodebottle-generative-ai#trial","runPath":"/mcp/github-genieincodebottle-generative-ai#trial","createEndpoint":"https://agentstaffer.com/api/actions/run-trial","createBody":{"listingId":"github:genieincodebottle/generative-ai","reviewer":"Trial reviewer","notes":"Sandbox run using mock data."}},"launchPacket":{"schemaVersion":"agentstaffer.trial-launch-packet-lite.v1","requiredPermissions":{"schemaVersion":"agentstaffer.trial-requirements.v1","principle":"Use the smallest sandbox permission set needed for the first reversible trial.","requirements":[{"id":"sandbox","label":"Sandbox workspace","detail":"Run the first trial in a throwaway directory or isolated environment.","severity":"required","text":"Sandbox workspace: Run the first trial in a throwaway directory or isolated environment."},{"id":"mcp-client","label":"MCP-capable client","detail":"Use Claude Desktop, Cursor, VS Code, or another client that can load MCP server configs.","severity":"required","text":"MCP-capable client: Use Claude Desktop, Cursor, VS Code, or another client that can load MCP server configs."},{"id":"temporary-credentials","label":"Temporary credentials","detail":"Use throwaway API keys or scoped test credentials. Do not start with production secrets.","severity":"conditional","text":"Temporary credentials: Use throwaway API keys or scoped test credentials. Do not start with production secrets."}],"humanReadable":["Sandbox workspace: Run the first trial in a throwaway directory or isolated environment.","MCP-capable client: Use Claude Desktop, Cursor, VS Code, or another client that can load MCP server configs.","Temporary credentials: Use throwaway API keys or scoped test credentials. Do not start with production secrets."]},"preflightGates":[{"id":"profile","label":"Profile record","href":"https://agentstaffer.com/mcp/github-genieincodebottle-generative-ai/profile.json","passed":true},{"id":"install","label":"Setup guide","href":"https://agentstaffer.com/mcp/github-genieincodebottle-generative-ai/install.json","passed":true},{"id":"sandbox","label":"Use sandbox data","passed":true}],"promotionGates":[{"id":"install","label":"Install path reviewed","passed":true},{"id":"trust","label":"Trust evidence reviewed","passed":false},{"id":"trial","label":"Sandbox trial recorded","passed":false}],"stopConditions":["Install asks for production credentials or broad write access.","Source or package cannot be verified from profile evidence."],"createRequest":{"endpoint":"https://agentstaffer.com/api/actions/run-trial","method":"POST","body":{"listingId":"github:genieincodebottle/generative-ai","reviewer":"Trial reviewer","notes":"Sandbox run using mock data."}}},"safety":{"trustTier":"Needs review","installReadiness":"MCP ready","topRisk":{"level":"medium","label":"No community evaluation","detail":"AgentStaffer users have not submitted a review yet."},"requirements":[{"id":"sandbox","label":"Sandbox workspace","detail":"Run the first trial in a throwaway directory or isolated environment.","severity":"required"},{"id":"mcp-client","label":"MCP-capable client","detail":"Use Claude Desktop, Cursor, VS Code, or another client that can load MCP server configs.","severity":"required"},{"id":"temporary-credentials","label":"Temporary credentials","detail":"Use throwaway API keys or scoped test credentials. Do not start with production secrets.","severity":"conditional"}],"permissionSummary":{"schemaVersion":"agentstaffer.trial-requirements.v1","principle":"Use the smallest sandbox permission set needed for the first reversible trial.","requirements":[{"id":"sandbox","label":"Sandbox workspace","detail":"Run the first trial in a throwaway directory or isolated environment.","severity":"required","text":"Sandbox workspace: Run the first trial in a throwaway directory or isolated environment."},{"id":"mcp-client","label":"MCP-capable client","detail":"Use Claude Desktop, Cursor, VS Code, or another client that can load MCP server configs.","severity":"required","text":"MCP-capable client: Use Claude Desktop, Cursor, VS Code, or another client that can load MCP server configs."},{"id":"temporary-credentials","label":"Temporary credentials","detail":"Use throwaway API keys or scoped test credentials. Do not start with production secrets.","severity":"conditional","text":"Temporary credentials: Use throwaway API keys or scoped test credentials. Do not start with production secrets."}],"humanReadable":["Sandbox workspace: Run the first trial in a throwaway directory or isolated environment.","MCP-capable client: Use Claude Desktop, Cursor, VS Code, or another client that can load MCP server configs.","Temporary credentials: Use throwaway API keys or scoped test credentials. Do not start with production secrets."]}}},{"id":"playbooks:skills-phrazzld-claude-config-llm-evaluation","slug":"playbooks-llm-evaluation-skill-by-phrazzld-claude-config","name":"llm-evaluation skill by phrazzld/claude-config","type":"skill","source":"playbooks","status":"pending","summary":"This skill enables automated LLM evaluation, regression and security testing with Promptfoo, integrating into CI/CD to improve prompt quality and safety.","profileUrl":"https://agentstaffer.com/skills/playbooks-llm-evaluation-skill-by-phrazzld-claude-config","fitScore":61,"roleFamily":"Workflow Automation","reviewSignal":"Pending","urls":{"profileJson":"https://agentstaffer.com/skills/playbooks-llm-evaluation-skill-by-phrazzld-claude-config/profile.json","installJson":"https://agentstaffer.com/skills/playbooks-llm-evaluation-skill-by-phrazzld-claude-config/install.json","trustJson":"https://agentstaffer.com/skills/playbooks-llm-evaluation-skill-by-phrazzld-claude-config/trust.json","trialJson":"https://agentstaffer.com/skills/playbooks-llm-evaluation-skill-by-phrazzld-claude-config/trial.json"},"queueRank":4,"recommendation":{"label":"Needs review","score":61},"trialPriority":81,"trial":{"title":"Run a dry-run workflow trial","input":"Connect the workflow to mock inputs and ask it to produce a trace without touching production systems.","expectedOutput":"A step-by-step run log with inputs, outputs, failures, and rollback notes.","safetyNotes":["Run in a sandbox or throwaway workspace.","Use non-production data and temporary credentials.","Capture output, logs, install friction, and any permission requests."],"decisionCriteria":["Shows observable steps","Handles failure paths","Documents required credentials","Can run without side effects"],"runUrl":"https://agentstaffer.com/skills/playbooks-llm-evaluation-skill-by-phrazzld-claude-config#trial","runPath":"/skills/playbooks-llm-evaluation-skill-by-phrazzld-claude-config#trial","createEndpoint":"https://agentstaffer.com/api/actions/run-trial","createBody":{"listingId":"playbooks:skills-phrazzld-claude-config-llm-evaluation","reviewer":"Trial reviewer","notes":"Sandbox run using mock data."}},"launchPacket":{"schemaVersion":"agentstaffer.trial-launch-packet-lite.v1","requiredPermissions":{"schemaVersion":"agentstaffer.trial-requirements.v1","principle":"Use the smallest sandbox permission set needed for the first reversible trial.","requirements":[{"id":"sandbox","label":"Sandbox workspace","detail":"Run the first trial in a throwaway directory or isolated environment.","severity":"required","text":"Sandbox workspace: Run the first trial in a throwaway directory or isolated environment."},{"id":"mcp-client","label":"MCP-capable client","detail":"Use Claude Desktop, Cursor, VS Code, or another client that can load MCP server configs.","severity":"required","text":"MCP-capable client: Use Claude Desktop, Cursor, VS Code, or another client that can load MCP server configs."},{"id":"terminal","label":"Local terminal","detail":"A CLI path is visible, but the exact command must be verified from source.","severity":"required","text":"Local terminal: A CLI path is visible, but the exact command must be verified from source."},{"id":"license-review","label":"License review","detail":"License metadata is missing from the current public profile.","severity":"review","text":"License review: License metadata is missing from the current public profile."}],"humanReadable":["Sandbox workspace: Run the first trial in a throwaway directory or isolated environment.","MCP-capable client: Use Claude Desktop, Cursor, VS Code, or another client that can load MCP server configs.","Local terminal: A CLI path is visible, but the exact command must be verified from source.","License review: License metadata is missing from the current public profile."]},"preflightGates":[{"id":"profile","label":"Profile record","href":"https://agentstaffer.com/skills/playbooks-llm-evaluation-skill-by-phrazzld-claude-config/profile.json","passed":true},{"id":"install","label":"Setup guide","href":"https://agentstaffer.com/skills/playbooks-llm-evaluation-skill-by-phrazzld-claude-config/install.json","passed":true},{"id":"sandbox","label":"Use sandbox data","passed":true}],"promotionGates":[{"id":"install","label":"Install path reviewed","passed":true},{"id":"trust","label":"Trust evidence reviewed","passed":false},{"id":"trial","label":"Sandbox trial recorded","passed":false}],"stopConditions":["Install asks for production credentials or broad write access.","Source or package cannot be verified from profile evidence."],"createRequest":{"endpoint":"https://agentstaffer.com/api/actions/run-trial","method":"POST","body":{"listingId":"playbooks:skills-phrazzld-claude-config-llm-evaluation","reviewer":"Trial reviewer","notes":"Sandbox run using mock data."}}},"safety":{"trustTier":"Needs review","installReadiness":"MCP ready","topRisk":{"level":"medium","label":"No community evaluation","detail":"AgentStaffer users have not submitted a review yet."},"requirements":[{"id":"sandbox","label":"Sandbox workspace","detail":"Run the first trial in a throwaway directory or isolated environment.","severity":"required"},{"id":"mcp-client","label":"MCP-capable client","detail":"Use Claude Desktop, Cursor, VS Code, or another client that can load MCP server configs.","severity":"required"},{"id":"terminal","label":"Local terminal","detail":"A CLI path is visible, but the exact command must be verified from source.","severity":"required"},{"id":"license-review","label":"License review","detail":"License metadata is missing from the current public profile.","severity":"review"}],"permissionSummary":{"schemaVersion":"agentstaffer.trial-requirements.v1","principle":"Use the smallest sandbox permission set needed for the first reversible trial.","requirements":[{"id":"sandbox","label":"Sandbox workspace","detail":"Run the first trial in a throwaway directory or isolated environment.","severity":"required","text":"Sandbox workspace: Run the first trial in a throwaway directory or isolated environment."},{"id":"mcp-client","label":"MCP-capable client","detail":"Use Claude Desktop, Cursor, VS Code, or another client that can load MCP server configs.","severity":"required","text":"MCP-capable client: Use Claude Desktop, Cursor, VS Code, or another client that can load MCP server configs."},{"id":"terminal","label":"Local terminal","detail":"A CLI path is visible, but the exact command must be verified from source.","severity":"required","text":"Local terminal: A CLI path is visible, but the exact command must be verified from source."},{"id":"license-review","label":"License review","detail":"License metadata is missing from the current public profile.","severity":"review","text":"License review: License metadata is missing from the current public profile."}],"humanReadable":["Sandbox workspace: Run the first trial in a throwaway directory or isolated environment.","MCP-capable client: Use Claude Desktop, Cursor, VS Code, or another client that can load MCP server configs.","Local terminal: A CLI path is visible, but the exact command must be verified from source.","License review: License metadata is missing from the current public profile."]}}},{"id":"playbooks:skills-xfstudio-skills-llm-evaluation","slug":"playbooks-llm-evaluation-skill-by-xfstudio-skills","name":"llm-evaluation skill by xfstudio/skills","type":"skill","source":"playbooks","status":"pending","summary":"This skill helps you evaluate LLM performance comprehensively using automated metrics, human feedback, and benchmarking to drive reliable improvements.","profileUrl":"https://agentstaffer.com/skills/playbooks-llm-evaluation-skill-by-xfstudio-skills","fitScore":61,"roleFamily":"Workflow Automation","reviewSignal":"Pending","urls":{"profileJson":"https://agentstaffer.com/skills/playbooks-llm-evaluation-skill-by-xfstudio-skills/profile.json","installJson":"https://agentstaffer.com/skills/playbooks-llm-evaluation-skill-by-xfstudio-skills/install.json","trustJson":"https://agentstaffer.com/skills/playbooks-llm-evaluation-skill-by-xfstudio-skills/trust.json","trialJson":"https://agentstaffer.com/skills/playbooks-llm-evaluation-skill-by-xfstudio-skills/trial.json"},"queueRank":5,"recommendation":{"label":"Needs review","score":61},"trialPriority":81,"trial":{"title":"Run a dry-run workflow trial","input":"Connect the workflow to mock inputs and ask it to produce a trace without touching production systems.","expectedOutput":"A step-by-step run log with inputs, outputs, failures, and rollback notes.","safetyNotes":["Run in a sandbox or throwaway workspace.","Use non-production data and temporary credentials.","Capture output, logs, install friction, and any permission requests."],"decisionCriteria":["Shows observable steps","Handles failure paths","Documents required credentials","Can run without side effects"],"runUrl":"https://agentstaffer.com/skills/playbooks-llm-evaluation-skill-by-xfstudio-skills#trial","runPath":"/skills/playbooks-llm-evaluation-skill-by-xfstudio-skills#trial","createEndpoint":"https://agentstaffer.com/api/actions/run-trial","createBody":{"listingId":"playbooks:skills-xfstudio-skills-llm-evaluation","reviewer":"Trial reviewer","notes":"Sandbox run using mock data."}},"launchPacket":{"schemaVersion":"agentstaffer.trial-launch-packet-lite.v1","requiredPermissions":{"schemaVersion":"agentstaffer.trial-requirements.v1","principle":"Use the smallest sandbox permission set needed for the first reversible trial.","requirements":[{"id":"sandbox","label":"Sandbox workspace","detail":"Run the first trial in a throwaway directory or isolated environment.","severity":"required","text":"Sandbox workspace: Run the first trial in a throwaway directory or isolated environment."},{"id":"mcp-client","label":"MCP-capable client","detail":"Use Claude Desktop, Cursor, VS Code, or another client that can load MCP server configs.","severity":"required","text":"MCP-capable client: Use Claude Desktop, Cursor, VS Code, or another client that can load MCP server configs."},{"id":"terminal","label":"Local terminal","detail":"A CLI path is visible, but the exact command must be verified from source.","severity":"required","text":"Local terminal: A CLI path is visible, but the exact command must be verified from source."},{"id":"license-review","label":"License review","detail":"License metadata is missing from the current public profile.","severity":"review","text":"License review: License metadata is missing from the current public profile."}],"humanReadable":["Sandbox workspace: Run the first trial in a throwaway directory or isolated environment.","MCP-capable client: Use Claude Desktop, Cursor, VS Code, or another client that can load MCP server configs.","Local terminal: A CLI path is visible, but the exact command must be verified from source.","License review: License metadata is missing from the current public profile."]},"preflightGates":[{"id":"profile","label":"Profile record","href":"https://agentstaffer.com/skills/playbooks-llm-evaluation-skill-by-xfstudio-skills/profile.json","passed":true},{"id":"install","label":"Setup guide","href":"https://agentstaffer.com/skills/playbooks-llm-evaluation-skill-by-xfstudio-skills/install.json","passed":true},{"id":"sandbox","label":"Use sandbox data","passed":true}],"promotionGates":[{"id":"install","label":"Install path reviewed","passed":true},{"id":"trust","label":"Trust evidence reviewed","passed":false},{"id":"trial","label":"Sandbox trial recorded","passed":false}],"stopConditions":["Install asks for production credentials or broad write access.","Source or package cannot be verified from profile evidence."],"createRequest":{"endpoint":"https://agentstaffer.com/api/actions/run-trial","method":"POST","body":{"listingId":"playbooks:skills-xfstudio-skills-llm-evaluation","reviewer":"Trial reviewer","notes":"Sandbox run using mock data."}}},"safety":{"trustTier":"Needs review","installReadiness":"MCP ready","topRisk":{"level":"medium","label":"No community evaluation","detail":"AgentStaffer users have not submitted a review yet."},"requirements":[{"id":"sandbox","label":"Sandbox workspace","detail":"Run the first trial in a throwaway directory or isolated environment.","severity":"required"},{"id":"mcp-client","label":"MCP-capable client","detail":"Use Claude Desktop, Cursor, VS Code, or another client that can load MCP server configs.","severity":"required"},{"id":"terminal","label":"Local terminal","detail":"A CLI path is visible, but the exact command must be verified from source.","severity":"required"},{"id":"license-review","label":"License review","detail":"License metadata is missing from the current public profile.","severity":"review"}],"permissionSummary":{"schemaVersion":"agentstaffer.trial-requirements.v1","principle":"Use the smallest sandbox permission set needed for the first reversible trial.","requirements":[{"id":"sandbox","label":"Sandbox workspace","detail":"Run the first trial in a throwaway directory or isolated environment.","severity":"required","text":"Sandbox workspace: Run the first trial in a throwaway directory or isolated environment."},{"id":"mcp-client","label":"MCP-capable client","detail":"Use Claude Desktop, Cursor, VS Code, or another client that can load MCP server configs.","severity":"required","text":"MCP-capable client: Use Claude Desktop, Cursor, VS Code, or another client that can load MCP server configs."},{"id":"terminal","label":"Local terminal","detail":"A CLI path is visible, but the exact command must be verified from source.","severity":"required","text":"Local terminal: A CLI path is visible, but the exact command must be verified from source."},{"id":"license-review","label":"License review","detail":"License metadata is missing from the current public profile.","severity":"review","text":"License review: License metadata is missing from the current public profile."}],"humanReadable":["Sandbox workspace: Run the first trial in a throwaway directory or isolated environment.","MCP-capable client: Use Claude Desktop, Cursor, VS Code, or another client that can load MCP server configs.","Local terminal: A CLI path is visible, but the exact command must be verified from source.","License review: License metadata is missing from the current public profile."]}}},{"id":"playbooks:skills-yonatangross-orchestkit-llm-evaluation","slug":"playbooks-llm-evaluation-skill-by-yonatangross-orchestkit","name":"llm-evaluation skill by yonatangross/orchestkit","type":"skill","source":"playbooks","status":"pending","summary":"This skill helps you evaluate and validate LLM outputs using multi-dimension scoring, gates, and hallucination checks for reliable AI pipelines.","profileUrl":"https://agentstaffer.com/skills/playbooks-llm-evaluation-skill-by-yonatangross-orchestkit","fitScore":55,"roleFamily":"Workflow Automation","reviewSignal":"Pending","urls":{"profileJson":"https://agentstaffer.com/skills/playbooks-llm-evaluation-skill-by-yonatangross-orchestkit/profile.json","installJson":"https://agentstaffer.com/skills/playbooks-llm-evaluation-skill-by-yonatangross-orchestkit/install.json","trustJson":"https://agentstaffer.com/skills/playbooks-llm-evaluation-skill-by-yonatangross-orchestkit/trust.json","trialJson":"https://agentstaffer.com/skills/playbooks-llm-evaluation-skill-by-yonatangross-orchestkit/trial.json"},"queueRank":6,"recommendation":{"label":"Needs review","score":55},"trialPriority":75,"trial":{"title":"Run a dry-run workflow trial","input":"Connect the workflow to mock inputs and ask it to produce a trace without touching production systems.","expectedOutput":"A step-by-step run log with inputs, outputs, failures, and rollback notes.","safetyNotes":["Run in a sandbox or throwaway workspace.","Use non-production data and temporary credentials.","Capture output, logs, install friction, and any permission requests."],"decisionCriteria":["Shows observable steps","Handles failure paths","Documents required credentials","Can run without side effects"],"runUrl":"https://agentstaffer.com/skills/playbooks-llm-evaluation-skill-by-yonatangross-orchestkit#trial","runPath":"/skills/playbooks-llm-evaluation-skill-by-yonatangross-orchestkit#trial","createEndpoint":"https://agentstaffer.com/api/actions/run-trial","createBody":{"listingId":"playbooks:skills-yonatangross-orchestkit-llm-evaluation","reviewer":"Trial reviewer","notes":"Sandbox run using mock data."}},"launchPacket":{"schemaVersion":"agentstaffer.trial-launch-packet-lite.v1","requiredPermissions":{"schemaVersion":"agentstaffer.trial-requirements.v1","principle":"Use the smallest sandbox permission set needed for the first reversible trial.","requirements":[{"id":"sandbox","label":"Sandbox workspace","detail":"Run the first trial in a throwaway directory or isolated environment.","severity":"required","text":"Sandbox workspace: Run the first trial in a throwaway directory or isolated environment."},{"id":"mcp-client","label":"MCP-capable client","detail":"Use Claude Desktop, Cursor, VS Code, or another client that can load MCP server configs.","severity":"required","text":"MCP-capable client: Use Claude Desktop, Cursor, VS Code, or another client that can load MCP server configs."},{"id":"terminal","label":"Local terminal","detail":"A CLI path is visible, but the exact command must be verified from source.","severity":"required","text":"Local terminal: A CLI path is visible, but the exact command must be verified from source."},{"id":"license-review","label":"License review","detail":"License metadata is missing from the current public profile.","severity":"review","text":"License review: License metadata is missing from the current public profile."}],"humanReadable":["Sandbox workspace: Run the first trial in a throwaway directory or isolated environment.","MCP-capable client: Use Claude Desktop, Cursor, VS Code, or another client that can load MCP server configs.","Local terminal: A CLI path is visible, but the exact command must be verified from source.","License review: License metadata is missing from the current public profile."]},"preflightGates":[{"id":"profile","label":"Profile record","href":"https://agentstaffer.com/skills/playbooks-llm-evaluation-skill-by-yonatangross-orchestkit/profile.json","passed":true},{"id":"install","label":"Setup guide","href":"https://agentstaffer.com/skills/playbooks-llm-evaluation-skill-by-yonatangross-orchestkit/install.json","passed":true},{"id":"sandbox","label":"Use sandbox data","passed":true}],"promotionGates":[{"id":"install","label":"Install path reviewed","passed":true},{"id":"trust","label":"Trust evidence reviewed","passed":false},{"id":"trial","label":"Sandbox trial recorded","passed":false}],"stopConditions":["Install asks for production credentials or broad write access.","Source or package cannot be verified from profile evidence."],"createRequest":{"endpoint":"https://agentstaffer.com/api/actions/run-trial","method":"POST","body":{"listingId":"playbooks:skills-yonatangross-orchestkit-llm-evaluation","reviewer":"Trial reviewer","notes":"Sandbox run using mock data."}}},"safety":{"trustTier":"Needs review","installReadiness":"MCP ready","topRisk":{"level":"medium","label":"No community evaluation","detail":"AgentStaffer users have not submitted a review yet."},"requirements":[{"id":"sandbox","label":"Sandbox workspace","detail":"Run the first trial in a throwaway directory or isolated environment.","severity":"required"},{"id":"mcp-client","label":"MCP-capable client","detail":"Use Claude Desktop, Cursor, VS Code, or another client that can load MCP server configs.","severity":"required"},{"id":"terminal","label":"Local terminal","detail":"A CLI path is visible, but the exact command must be verified from source.","severity":"required"},{"id":"license-review","label":"License review","detail":"License metadata is missing from the current public profile.","severity":"review"}],"permissionSummary":{"schemaVersion":"agentstaffer.trial-requirements.v1","principle":"Use the smallest sandbox permission set needed for the first reversible trial.","requirements":[{"id":"sandbox","label":"Sandbox workspace","detail":"Run the first trial in a throwaway directory or isolated environment.","severity":"required","text":"Sandbox workspace: Run the first trial in a throwaway directory or isolated environment."},{"id":"mcp-client","label":"MCP-capable client","detail":"Use Claude Desktop, Cursor, VS Code, or another client that can load MCP server configs.","severity":"required","text":"MCP-capable client: Use Claude Desktop, Cursor, VS Code, or another client that can load MCP server configs."},{"id":"terminal","label":"Local terminal","detail":"A CLI path is visible, but the exact command must be verified from source.","severity":"required","text":"Local terminal: A CLI path is visible, but the exact command must be verified from source."},{"id":"license-review","label":"License review","detail":"License metadata is missing from the current public profile.","severity":"review","text":"License review: License metadata is missing from the current public profile."}],"humanReadable":["Sandbox workspace: Run the first trial in a throwaway directory or isolated environment.","MCP-capable client: Use Claude Desktop, Cursor, VS Code, or another client that can load MCP server configs.","Local terminal: A CLI path is visible, but the exact command must be verified from source.","License review: License metadata is missing from the current public profile."]}}},{"id":"mcpservers:agent-skills-posthog-exploring-llm-evaluations","slug":"mcpservers-exploring-llm-evaluations","name":"exploring-llm-evaluations","type":"skill","source":"mcpservers","status":"pending","summary":"PostHog evaluations score $ai_generation events. Each evaluation is one of two types, both first-class:","profileUrl":"https://agentstaffer.com/skills/mcpservers-exploring-llm-evaluations","fitScore":61,"roleFamily":"Coding & Development","reviewSignal":"Pending","urls":{"profileJson":"https://agentstaffer.com/skills/mcpservers-exploring-llm-evaluations/profile.json","installJson":"https://agentstaffer.com/skills/mcpservers-exploring-llm-evaluations/install.json","trustJson":"https://agentstaffer.com/skills/mcpservers-exploring-llm-evaluations/trust.json","trialJson":"https://agentstaffer.com/skills/mcpservers-exploring-llm-evaluations/trial.json"},"queueRank":7,"recommendation":{"label":"Needs review","score":61},"trialPriority":81,"trial":{"title":"Run a small code review trial","input":"Give the agent one low-risk issue or one small pull request and ask for defects, tests, and a patch plan.","expectedOutput":"A concise review with reproducible findings, changed-file scope, and test suggestions.","safetyNotes":["Run in a sandbox or throwaway workspace.","Use non-production data and temporary credentials.","Capture output, logs, install friction, and any permission requests."],"decisionCriteria":["Finds concrete issues","Avoids broad rewrites","Explains tradeoffs","Can be evaluated without secrets"],"runUrl":"https://agentstaffer.com/skills/mcpservers-exploring-llm-evaluations#trial","runPath":"/skills/mcpservers-exploring-llm-evaluations#trial","createEndpoint":"https://agentstaffer.com/api/actions/run-trial","createBody":{"listingId":"mcpservers:agent-skills-posthog-exploring-llm-evaluations","reviewer":"Trial reviewer","notes":"Sandbox run using mock data."}},"launchPacket":{"schemaVersion":"agentstaffer.trial-launch-packet-lite.v1","requiredPermissions":{"schemaVersion":"agentstaffer.trial-requirements.v1","principle":"Use the smallest sandbox permission set needed for the first reversible trial.","requirements":[{"id":"sandbox","label":"Sandbox workspace","detail":"Run the first trial in a throwaway directory or isolated environment.","severity":"required","text":"Sandbox workspace: Run the first trial in a throwaway directory or isolated environment."},{"id":"mcp-client","label":"MCP-capable client","detail":"Use Claude Desktop, Cursor, VS Code, or another client that can load MCP server configs.","severity":"required","text":"MCP-capable client: Use Claude Desktop, Cursor, VS Code, or another client that can load MCP server configs."},{"id":"terminal","label":"Local terminal","detail":"A CLI path is visible, but the exact command must be verified from source.","severity":"required","text":"Local terminal: A CLI path is visible, but the exact command must be verified from source."},{"id":"license-review","label":"License review","detail":"License metadata is missing from the current public profile.","severity":"review","text":"License review: License metadata is missing from the current public profile."}],"humanReadable":["Sandbox workspace: Run the first trial in a throwaway directory or isolated environment.","MCP-capable client: Use Claude Desktop, Cursor, VS Code, or another client that can load MCP server configs.","Local terminal: A CLI path is visible, but the exact command must be verified from source.","License review: License metadata is missing from the current public profile."]},"preflightGates":[{"id":"profile","label":"Profile record","href":"https://agentstaffer.com/skills/mcpservers-exploring-llm-evaluations/profile.json","passed":true},{"id":"install","label":"Setup guide","href":"https://agentstaffer.com/skills/mcpservers-exploring-llm-evaluations/install.json","passed":true},{"id":"sandbox","label":"Use sandbox data","passed":true}],"promotionGates":[{"id":"install","label":"Install path reviewed","passed":true},{"id":"trust","label":"Trust evidence reviewed","passed":false},{"id":"trial","label":"Sandbox trial recorded","passed":false}],"stopConditions":["Install asks for production credentials or broad write access.","Source or package cannot be verified from profile evidence."],"createRequest":{"endpoint":"https://agentstaffer.com/api/actions/run-trial","method":"POST","body":{"listingId":"mcpservers:agent-skills-posthog-exploring-llm-evaluations","reviewer":"Trial reviewer","notes":"Sandbox run using mock data."}}},"safety":{"trustTier":"Needs review","installReadiness":"MCP ready","topRisk":{"level":"medium","label":"No community evaluation","detail":"AgentStaffer users have not submitted a review yet."},"requirements":[{"id":"sandbox","label":"Sandbox workspace","detail":"Run the first trial in a throwaway directory or isolated environment.","severity":"required"},{"id":"mcp-client","label":"MCP-capable client","detail":"Use Claude Desktop, Cursor, VS Code, or another client that can load MCP server configs.","severity":"required"},{"id":"terminal","label":"Local terminal","detail":"A CLI path is visible, but the exact command must be verified from source.","severity":"required"},{"id":"license-review","label":"License review","detail":"License metadata is missing from the current public profile.","severity":"review"}],"permissionSummary":{"schemaVersion":"agentstaffer.trial-requirements.v1","principle":"Use the smallest sandbox permission set needed for the first reversible trial.","requirements":[{"id":"sandbox","label":"Sandbox workspace","detail":"Run the first trial in a throwaway directory or isolated environment.","severity":"required","text":"Sandbox workspace: Run the first trial in a throwaway directory or isolated environment."},{"id":"mcp-client","label":"MCP-capable client","detail":"Use Claude Desktop, Cursor, VS Code, or another client that can load MCP server configs.","severity":"required","text":"MCP-capable client: Use Claude Desktop, Cursor, VS Code, or another client that can load MCP server configs."},{"id":"terminal","label":"Local terminal","detail":"A CLI path is visible, but the exact command must be verified from source.","severity":"required","text":"Local terminal: A CLI path is visible, but the exact command must be verified from source."},{"id":"license-review","label":"License review","detail":"License metadata is missing from the current public profile.","severity":"review","text":"License review: License metadata is missing from the current public profile."}],"humanReadable":["Sandbox workspace: Run the first trial in a throwaway directory or isolated environment.","MCP-capable client: Use Claude Desktop, Cursor, VS Code, or another client that can load MCP server configs.","Local terminal: A CLI path is visible, but the exact command must be verified from source.","License review: License metadata is missing from the current public profile."]}}},{"id":"remoteopenclaw:skills-wshobson-agents-llm-evaluation","slug":"remoteopenclaw-llm-evaluation","name":"llm-evaluation","type":"skill","source":"remoteopenclaw","status":"pending","summary":"Implement comprehensive evaluation strategies for LLM applications using automated metrics, human feedback, and benchmarking. Use when testing LLM performance, measuring AI application quality, or establishing evaluation frameworks.","profileUrl":"https://agentstaffer.com/skills/remoteopenclaw-llm-evaluation","fitScore":44,"roleFamily":"Coding & Development","reviewSignal":"Pending","urls":{"profileJson":"https://agentstaffer.com/skills/remoteopenclaw-llm-evaluation/profile.json","installJson":"https://agentstaffer.com/skills/remoteopenclaw-llm-evaluation/install.json","trustJson":"https://agentstaffer.com/skills/remoteopenclaw-llm-evaluation/trust.json","trialJson":"https://agentstaffer.com/skills/remoteopenclaw-llm-evaluation/trial.json"},"queueRank":8,"recommendation":{"label":"Needs review","score":44},"trialPriority":64,"trial":{"title":"Run a small code review trial","input":"Give the agent one low-risk issue or one small pull request and ask for defects, tests, and a patch plan.","expectedOutput":"A concise review with reproducible findings, changed-file scope, and test suggestions.","safetyNotes":["Run in a sandbox or throwaway workspace.","Use non-production data and temporary credentials.","Capture output, logs, install friction, and any permission requests."],"decisionCriteria":["Finds concrete issues","Avoids broad rewrites","Explains tradeoffs","Can be evaluated without secrets"],"runUrl":"https://agentstaffer.com/skills/remoteopenclaw-llm-evaluation#trial","runPath":"/skills/remoteopenclaw-llm-evaluation#trial","createEndpoint":"https://agentstaffer.com/api/actions/run-trial","createBody":{"listingId":"remoteopenclaw:skills-wshobson-agents-llm-evaluation","reviewer":"Trial reviewer","notes":"Sandbox run using mock data."}},"launchPacket":{"schemaVersion":"agentstaffer.trial-launch-packet-lite.v1","requiredPermissions":{"schemaVersion":"agentstaffer.trial-requirements.v1","principle":"Use the smallest sandbox permission set needed for the first reversible trial.","requirements":[{"id":"sandbox","label":"Sandbox workspace","detail":"Run the first trial in a throwaway directory or isolated environment.","severity":"required","text":"Sandbox workspace: Run the first trial in a throwaway directory or isolated environment."},{"id":"license-review","label":"License review","detail":"License metadata is missing from the current public profile.","severity":"review","text":"License review: License metadata is missing from the current public profile."}],"humanReadable":["Sandbox workspace: Run the first trial in a throwaway directory or isolated environment.","License review: License metadata is missing from the current public profile."]},"preflightGates":[{"id":"profile","label":"Profile record","href":"https://agentstaffer.com/skills/remoteopenclaw-llm-evaluation/profile.json","passed":true},{"id":"install","label":"Setup guide","href":"https://agentstaffer.com/skills/remoteopenclaw-llm-evaluation/install.json","passed":true},{"id":"sandbox","label":"Use sandbox data","passed":true}],"promotionGates":[{"id":"install","label":"Install path reviewed","passed":true},{"id":"trust","label":"Trust evidence reviewed","passed":false},{"id":"trial","label":"Sandbox trial recorded","passed":false}],"stopConditions":["Install asks for production credentials or broad write access.","Source or package cannot be verified from profile evidence."],"createRequest":{"endpoint":"https://agentstaffer.com/api/actions/run-trial","method":"POST","body":{"listingId":"remoteopenclaw:skills-wshobson-agents-llm-evaluation","reviewer":"Trial reviewer","notes":"Sandbox run using mock data."}}},"safety":{"trustTier":"Needs review","installReadiness":"Source guide","topRisk":{"level":"medium","label":"No community evaluation","detail":"AgentStaffer users have not submitted a review yet."},"requirements":[{"id":"sandbox","label":"Sandbox workspace","detail":"Run the first trial in a throwaway directory or isolated environment.","severity":"required"},{"id":"license-review","label":"License review","detail":"License metadata is missing from the current public profile.","severity":"review"}],"permissionSummary":{"schemaVersion":"agentstaffer.trial-requirements.v1","principle":"Use the smallest sandbox permission set needed for the first reversible trial.","requirements":[{"id":"sandbox","label":"Sandbox workspace","detail":"Run the first trial in a throwaway directory or isolated environment.","severity":"required","text":"Sandbox workspace: Run the first trial in a throwaway directory or isolated environment."},{"id":"license-review","label":"License review","detail":"License metadata is missing from the current public profile.","severity":"review","text":"License review: License metadata is missing from the current public profile."}],"humanReadable":["Sandbox workspace: Run the first trial in a throwaway directory or isolated environment.","License review: License metadata is missing from the current public profile."]}}},{"id":"enterprisedna:directories-open-source-autorag","slug":"enterprisedna-autorag","name":"AutoRAG","type":"framework","source":"enterprisedna","status":"pending","summary":"AutoRAG is an open-source framework that automates the evaluation and optimization of retrieval-augmented generation pipelines. It uses an AutoML-style approach to test multiple configurations of retrievers, generators, and other components and identifies the best combination for a given dataset. Users provide a dataset and a set of candidate modules, and A...","profileUrl":"https://agentstaffer.com/mcp/enterprisedna-autorag","fitScore":52,"roleFamily":"Coding & Development","reviewSignal":"Pending","urls":{"profileJson":"https://agentstaffer.com/mcp/enterprisedna-autorag/profile.json","installJson":"https://agentstaffer.com/mcp/enterprisedna-autorag/install.json","trustJson":"https://agentstaffer.com/mcp/enterprisedna-autorag/trust.json","trialJson":"https://agentstaffer.com/mcp/enterprisedna-autorag/trial.json"},"queueRank":9,"recommendation":{"label":"Needs review","score":52},"trialPriority":72,"trial":{"title":"Run a small code review trial","input":"Give the agent one low-risk issue or one small pull request and ask for defects, tests, and a patch plan.","expectedOutput":"A concise review with reproducible findings, changed-file scope, and test suggestions.","safetyNotes":["Run in a sandbox or throwaway workspace.","Use non-production data and temporary credentials.","Capture output, logs, install friction, and any permission requests."],"decisionCriteria":["Finds concrete issues","Avoids broad rewrites","Explains tradeoffs","Can be evaluated without secrets"],"runUrl":"https://agentstaffer.com/mcp/enterprisedna-autorag#trial","runPath":"/mcp/enterprisedna-autorag#trial","createEndpoint":"https://agentstaffer.com/api/actions/run-trial","createBody":{"listingId":"enterprisedna:directories-open-source-autorag","reviewer":"Trial reviewer","notes":"Sandbox run using mock data."}},"launchPacket":{"schemaVersion":"agentstaffer.trial-launch-packet-lite.v1","requiredPermissions":{"schemaVersion":"agentstaffer.trial-requirements.v1","principle":"Use the smallest sandbox permission set needed for the first reversible trial.","requirements":[{"id":"sandbox","label":"Sandbox workspace","detail":"Run the first trial in a throwaway directory or isolated environment.","severity":"required","text":"Sandbox workspace: Run the first trial in a throwaway directory or isolated environment."},{"id":"mcp-client","label":"MCP-capable client","detail":"Use Claude Desktop, Cursor, VS Code, or another client that can load MCP server configs.","severity":"required","text":"MCP-capable client: Use Claude Desktop, Cursor, VS Code, or another client that can load MCP server configs."},{"id":"license-review","label":"License review","detail":"License metadata is missing from the current public profile.","severity":"review","text":"License review: License metadata is missing from the current public profile."}],"humanReadable":["Sandbox workspace: Run the first trial in a throwaway directory or isolated environment.","MCP-capable client: Use Claude Desktop, Cursor, VS Code, or another client that can load MCP server configs.","License review: License metadata is missing from the current public profile."]},"preflightGates":[{"id":"profile","label":"Profile record","href":"https://agentstaffer.com/mcp/enterprisedna-autorag/profile.json","passed":true},{"id":"install","label":"Setup guide","href":"https://agentstaffer.com/mcp/enterprisedna-autorag/install.json","passed":true},{"id":"sandbox","label":"Use sandbox data","passed":true}],"promotionGates":[{"id":"install","label":"Install path reviewed","passed":true},{"id":"trust","label":"Trust evidence reviewed","passed":false},{"id":"trial","label":"Sandbox trial recorded","passed":false}],"stopConditions":["Install asks for production credentials or broad write access.","Source or package cannot be verified from profile evidence."],"createRequest":{"endpoint":"https://agentstaffer.com/api/actions/run-trial","method":"POST","body":{"listingId":"enterprisedna:directories-open-source-autorag","reviewer":"Trial reviewer","notes":"Sandbox run using mock data."}}},"safety":{"trustTier":"Needs review","installReadiness":"MCP ready","topRisk":{"level":"medium","label":"No community evaluation","detail":"AgentStaffer users have not submitted a review yet."},"requirements":[{"id":"sandbox","label":"Sandbox workspace","detail":"Run the first trial in a throwaway directory or isolated environment.","severity":"required"},{"id":"mcp-client","label":"MCP-capable client","detail":"Use Claude Desktop, Cursor, VS Code, or another client that can load MCP server configs.","severity":"required"},{"id":"license-review","label":"License review","detail":"License metadata is missing from the current public profile.","severity":"review"}],"permissionSummary":{"schemaVersion":"agentstaffer.trial-requirements.v1","principle":"Use the smallest sandbox permission set needed for the first reversible trial.","requirements":[{"id":"sandbox","label":"Sandbox workspace","detail":"Run the first trial in a throwaway directory or isolated environment.","severity":"required","text":"Sandbox workspace: Run the first trial in a throwaway directory or isolated environment."},{"id":"mcp-client","label":"MCP-capable client","detail":"Use Claude Desktop, Cursor, VS Code, or another client that can load MCP server configs.","severity":"required","text":"MCP-capable client: Use Claude Desktop, Cursor, VS Code, or another client that can load MCP server configs."},{"id":"license-review","label":"License review","detail":"License metadata is missing from the current public profile.","severity":"review","text":"License review: License metadata is missing from the current public profile."}],"humanReadable":["Sandbox workspace: Run the first trial in a throwaway directory or isolated environment.","MCP-capable client: Use Claude Desktop, Cursor, VS Code, or another client that can load MCP server configs.","License review: License metadata is missing from the current public profile."]}}},{"id":"github:jeinlee1991/chinese-llm-benchmark","slug":"github-jeinlee1991-chinese-llm-benchmark","name":"jeinlee1991/chinese-llm-benchmark","type":"model","source":"github","status":"pending","summary":"Chinese LLM benchmark and defect database covering hundreds of commercial and open-source models with ranking and analysis data.","profileUrl":"https://agentstaffer.com/mcp/github-jeinlee1991-chinese-llm-benchmark","fitScore":64,"roleFamily":"Coding & Development","reviewSignal":"Pending","urls":{"profileJson":"https://agentstaffer.com/mcp/github-jeinlee1991-chinese-llm-benchmark/profile.json","installJson":"https://agentstaffer.com/mcp/github-jeinlee1991-chinese-llm-benchmark/install.json","trustJson":"https://agentstaffer.com/mcp/github-jeinlee1991-chinese-llm-benchmark/trust.json","trialJson":"https://agentstaffer.com/mcp/github-jeinlee1991-chinese-llm-benchmark/trial.json"},"queueRank":10,"recommendation":{"label":"Needs review","score":64},"trialPriority":84,"trial":{"title":"Run a small code review trial","input":"Give the agent one low-risk issue or one small pull request and ask for defects, tests, and a patch plan.","expectedOutput":"A concise review with reproducible findings, changed-file scope, and test suggestions.","safetyNotes":["Run in a sandbox or throwaway workspace.","Use non-production data and temporary credentials.","Capture output, logs, install friction, and any permission requests."],"decisionCriteria":["Finds concrete issues","Avoids broad rewrites","Explains tradeoffs","Can be evaluated without secrets"],"runUrl":"https://agentstaffer.com/mcp/github-jeinlee1991-chinese-llm-benchmark#trial","runPath":"/mcp/github-jeinlee1991-chinese-llm-benchmark#trial","createEndpoint":"https://agentstaffer.com/api/actions/run-trial","createBody":{"listingId":"github:jeinlee1991/chinese-llm-benchmark","reviewer":"Trial reviewer","notes":"Sandbox run using mock data."}},"launchPacket":{"schemaVersion":"agentstaffer.trial-launch-packet-lite.v1","requiredPermissions":{"schemaVersion":"agentstaffer.trial-requirements.v1","principle":"Use the smallest sandbox permission set needed for the first reversible trial.","requirements":[{"id":"sandbox","label":"Sandbox workspace","detail":"Run the first trial in a throwaway directory or isolated environment.","severity":"required","text":"Sandbox workspace: Run the first trial in a throwaway directory or isolated environment."},{"id":"mcp-client","label":"MCP-capable client","detail":"Use Claude Desktop, Cursor, VS Code, or another client that can load MCP server configs.","severity":"required","text":"MCP-capable client: Use Claude Desktop, Cursor, VS Code, or another client that can load MCP server configs."},{"id":"license-review","label":"License review","detail":"License metadata is missing from the current public profile.","severity":"review","text":"License review: License metadata is missing from the current public profile."}],"humanReadable":["Sandbox workspace: Run the first trial in a throwaway directory or isolated environment.","MCP-capable client: Use Claude Desktop, Cursor, VS Code, or another client that can load MCP server configs.","License review: License metadata is missing from the current public profile."]},"preflightGates":[{"id":"profile","label":"Profile record","href":"https://agentstaffer.com/mcp/github-jeinlee1991-chinese-llm-benchmark/profile.json","passed":true},{"id":"install","label":"Setup guide","href":"https://agentstaffer.com/mcp/github-jeinlee1991-chinese-llm-benchmark/install.json","passed":true},{"id":"sandbox","label":"Use sandbox data","passed":true}],"promotionGates":[{"id":"install","label":"Install path reviewed","passed":true},{"id":"trust","label":"Trust evidence reviewed","passed":false},{"id":"trial","label":"Sandbox trial recorded","passed":false}],"stopConditions":["Install asks for production credentials or broad write access.","Source or package cannot be verified from profile evidence."],"createRequest":{"endpoint":"https://agentstaffer.com/api/actions/run-trial","method":"POST","body":{"listingId":"github:jeinlee1991/chinese-llm-benchmark","reviewer":"Trial reviewer","notes":"Sandbox run using mock data."}}},"safety":{"trustTier":"Needs review","installReadiness":"MCP ready","topRisk":{"level":"medium","label":"No community evaluation","detail":"AgentStaffer users have not submitted a review yet."},"requirements":[{"id":"sandbox","label":"Sandbox workspace","detail":"Run the first trial in a throwaway directory or isolated environment.","severity":"required"},{"id":"mcp-client","label":"MCP-capable client","detail":"Use Claude Desktop, Cursor, VS Code, or another client that can load MCP server configs.","severity":"required"},{"id":"license-review","label":"License review","detail":"License metadata is missing from the current public profile.","severity":"review"}],"permissionSummary":{"schemaVersion":"agentstaffer.trial-requirements.v1","principle":"Use the smallest sandbox permission set needed for the first reversible trial.","requirements":[{"id":"sandbox","label":"Sandbox workspace","detail":"Run the first trial in a throwaway directory or isolated environment.","severity":"required","text":"Sandbox workspace: Run the first trial in a throwaway directory or isolated environment."},{"id":"mcp-client","label":"MCP-capable client","detail":"Use Claude Desktop, Cursor, VS Code, or another client that can load MCP server configs.","severity":"required","text":"MCP-capable client: Use Claude Desktop, Cursor, VS Code, or another client that can load MCP server configs."},{"id":"license-review","label":"License review","detail":"License metadata is missing from the current public profile.","severity":"review","text":"License review: License metadata is missing from the current public profile."}],"humanReadable":["Sandbox workspace: Run the first trial in a throwaway directory or isolated environment.","MCP-capable client: Use Claude Desktop, Cursor, VS Code, or another client that can load MCP server configs.","License review: License metadata is missing from the current public profile."]}}},{"id":"agensi:skills-ai-feature-eval-writer-golden-datasets-rubrics-and-llm-as-judge-prompts-that-actually-catch-regressions","slug":"agensi-ai-feature-eval-writer","name":"AI Feature Eval Writer","type":"skill","source":"agensi","status":"pending","summary":"Design and write the eval suite for your LLM-powered feature — the metrics that match your failure modes, a golden dataset plan with starter cases, anchored rubrics, LLM-as-judge prompts with the known bias mitigations, and pass/fail gates wired for CI.","profileUrl":"https://agentstaffer.com/skills/agensi-ai-feature-eval-writer","fitScore":39,"roleFamily":"Coding & Development","reviewSignal":"Pending","urls":{"profileJson":"https://agentstaffer.com/skills/agensi-ai-feature-eval-writer/profile.json","installJson":"https://agentstaffer.com/skills/agensi-ai-feature-eval-writer/install.json","trustJson":"https://agentstaffer.com/skills/agensi-ai-feature-eval-writer/trust.json","trialJson":"https://agentstaffer.com/skills/agensi-ai-feature-eval-writer/trial.json"},"queueRank":11,"recommendation":{"label":"Needs review","score":39},"trialPriority":59,"trial":{"title":"Run a small code review trial","input":"Give the agent one low-risk issue or one small pull request and ask for defects, tests, and a patch plan.","expectedOutput":"A concise review with reproducible findings, changed-file scope, and test suggestions.","safetyNotes":["Run in a sandbox or throwaway workspace.","Use non-production data and temporary credentials.","Capture output, logs, install friction, and any permission requests."],"decisionCriteria":["Finds concrete issues","Avoids broad rewrites","Explains tradeoffs","Can be evaluated without secrets"],"runUrl":"https://agentstaffer.com/skills/agensi-ai-feature-eval-writer#trial","runPath":"/skills/agensi-ai-feature-eval-writer#trial","createEndpoint":"https://agentstaffer.com/api/actions/run-trial","createBody":{"listingId":"agensi:skills-ai-feature-eval-writer-golden-datasets-rubrics-and-llm-as-judge-prompts-that-actually-catch-regressions","reviewer":"Trial reviewer","notes":"Sandbox run using mock data."}},"launchPacket":{"schemaVersion":"agentstaffer.trial-launch-packet-lite.v1","requiredPermissions":{"schemaVersion":"agentstaffer.trial-requirements.v1","principle":"Use the smallest sandbox permission set needed for the first reversible trial.","requirements":[{"id":"sandbox","label":"Sandbox workspace","detail":"Run the first trial in a throwaway directory or isolated environment.","severity":"required","text":"Sandbox workspace: Run the first trial in a throwaway directory or isolated environment."},{"id":"license-review","label":"License review","detail":"License metadata is missing from the current public profile.","severity":"review","text":"License review: License metadata is missing from the current public profile."}],"humanReadable":["Sandbox workspace: Run the first trial in a throwaway directory or isolated environment.","License review: License metadata is missing from the current public profile."]},"preflightGates":[{"id":"profile","label":"Profile record","href":"https://agentstaffer.com/skills/agensi-ai-feature-eval-writer/profile.json","passed":true},{"id":"install","label":"Setup guide","href":"https://agentstaffer.com/skills/agensi-ai-feature-eval-writer/install.json","passed":true},{"id":"sandbox","label":"Use sandbox data","passed":true}],"promotionGates":[{"id":"install","label":"Install path reviewed","passed":true},{"id":"trust","label":"Trust evidence reviewed","passed":true},{"id":"trial","label":"Sandbox trial recorded","passed":false}],"stopConditions":["Install asks for production credentials or broad write access.","Source or package cannot be verified from profile evidence."],"createRequest":{"endpoint":"https://agentstaffer.com/api/actions/run-trial","method":"POST","body":{"listingId":"agensi:skills-ai-feature-eval-writer-golden-datasets-rubrics-and-llm-as-judge-prompts-that-actually-catch-regressions","reviewer":"Trial reviewer","notes":"Sandbox run using mock data."}}},"safety":{"trustTier":"Verified source","installReadiness":"Source guide","topRisk":{"level":"medium","label":"No community evaluation","detail":"AgentStaffer users have not submitted a review yet."},"requirements":[{"id":"sandbox","label":"Sandbox workspace","detail":"Run the first trial in a throwaway directory or isolated environment.","severity":"required"},{"id":"license-review","label":"License review","detail":"License metadata is missing from the current public profile.","severity":"review"}],"permissionSummary":{"schemaVersion":"agentstaffer.trial-requirements.v1","principle":"Use the smallest sandbox permission set needed for the first reversible trial.","requirements":[{"id":"sandbox","label":"Sandbox workspace","detail":"Run the first trial in a throwaway directory or isolated environment.","severity":"required","text":"Sandbox workspace: Run the first trial in a throwaway directory or isolated environment."},{"id":"license-review","label":"License review","detail":"License metadata is missing from the current public profile.","severity":"review","text":"License review: License metadata is missing from the current public profile."}],"humanReadable":["Sandbox workspace: Run the first trial in a throwaway directory or isolated environment.","License review: License metadata is missing from the current public profile."]}}},{"id":"remoteopenclaw:skills-hermes-community-wshobson-agents-llm-evaluation","slug":"remoteopenclaw-wshobson-agents-llm-evaluation","name":"Wshobson Agents LLM Evaluation","type":"skill","source":"remoteopenclaw","status":"pending","summary":"Community skill discovered through the LobeHub marketplace. Open the original listing for install instructions, files, and updates.","profileUrl":"https://agentstaffer.com/skills/remoteopenclaw-wshobson-agents-llm-evaluation","fitScore":35,"roleFamily":"Workflow Automation","reviewSignal":"Pending","urls":{"profileJson":"https://agentstaffer.com/skills/remoteopenclaw-wshobson-agents-llm-evaluation/profile.json","installJson":"https://agentstaffer.com/skills/remoteopenclaw-wshobson-agents-llm-evaluation/install.json","trustJson":"https://agentstaffer.com/skills/remoteopenclaw-wshobson-agents-llm-evaluation/trust.json","trialJson":"https://agentstaffer.com/skills/remoteopenclaw-wshobson-agents-llm-evaluation/trial.json"},"queueRank":12,"recommendation":{"label":"Needs review","score":35},"trialPriority":55,"trial":{"title":"Run a dry-run workflow trial","input":"Connect the workflow to mock inputs and ask it to produce a trace without touching production systems.","expectedOutput":"A step-by-step run log with inputs, outputs, failures, and rollback notes.","safetyNotes":["Run in a sandbox or throwaway workspace.","Use non-production data and temporary credentials.","Capture output, logs, install friction, and any permission requests."],"decisionCriteria":["Shows observable steps","Handles failure paths","Documents required credentials","Can run without side effects"],"runUrl":"https://agentstaffer.com/skills/remoteopenclaw-wshobson-agents-llm-evaluation#trial","runPath":"/skills/remoteopenclaw-wshobson-agents-llm-evaluation#trial","createEndpoint":"https://agentstaffer.com/api/actions/run-trial","createBody":{"listingId":"remoteopenclaw:skills-hermes-community-wshobson-agents-llm-evaluation","reviewer":"Trial reviewer","notes":"Sandbox run using mock data."}},"launchPacket":{"schemaVersion":"agentstaffer.trial-launch-packet-lite.v1","requiredPermissions":{"schemaVersion":"agentstaffer.trial-requirements.v1","principle":"Use the smallest sandbox permission set needed for the first reversible trial.","requirements":[{"id":"sandbox","label":"Sandbox workspace","detail":"Run the first trial in a throwaway directory or isolated environment.","severity":"required","text":"Sandbox workspace: Run the first trial in a throwaway directory or isolated environment."},{"id":"license-review","label":"License review","detail":"License metadata is missing from the current public profile.","severity":"review","text":"License review: License metadata is missing from the current public profile."}],"humanReadable":["Sandbox workspace: Run the first trial in a throwaway directory or isolated environment.","License review: License metadata is missing from the current public profile."]},"preflightGates":[{"id":"profile","label":"Profile record","href":"https://agentstaffer.com/skills/remoteopenclaw-wshobson-agents-llm-evaluation/profile.json","passed":true},{"id":"install","label":"Setup guide","href":"https://agentstaffer.com/skills/remoteopenclaw-wshobson-agents-llm-evaluation/install.json","passed":true},{"id":"sandbox","label":"Use sandbox data","passed":true}],"promotionGates":[{"id":"install","label":"Install path reviewed","passed":true},{"id":"trust","label":"Trust evidence reviewed","passed":false},{"id":"trial","label":"Sandbox trial recorded","passed":false}],"stopConditions":["Install asks for production credentials or broad write access.","Source or package cannot be verified from profile evidence."],"createRequest":{"endpoint":"https://agentstaffer.com/api/actions/run-trial","method":"POST","body":{"listingId":"remoteopenclaw:skills-hermes-community-wshobson-agents-llm-evaluation","reviewer":"Trial reviewer","notes":"Sandbox run using mock data."}}},"safety":{"trustTier":"Needs review","installReadiness":"Source guide","topRisk":{"level":"medium","label":"No community evaluation","detail":"AgentStaffer users have not submitted a review yet."},"requirements":[{"id":"sandbox","label":"Sandbox workspace","detail":"Run the first trial in a throwaway directory or isolated environment.","severity":"required"},{"id":"license-review","label":"License review","detail":"License metadata is missing from the current public profile.","severity":"review"}],"permissionSummary":{"schemaVersion":"agentstaffer.trial-requirements.v1","principle":"Use the smallest sandbox permission set needed for the first reversible trial.","requirements":[{"id":"sandbox","label":"Sandbox workspace","detail":"Run the first trial in a throwaway directory or isolated environment.","severity":"required","text":"Sandbox workspace: Run the first trial in a throwaway directory or isolated environment."},{"id":"license-review","label":"License review","detail":"License metadata is missing from the current public profile.","severity":"review","text":"License review: License metadata is missing from the current public profile."}],"humanReadable":["Sandbox workspace: Run the first trial in a throwaway directory or isolated environment.","License review: License metadata is missing from the current public profile."]}}},{"id":"claudemarketplaces:skills-sickn33-antigravity-awesome-skills-llm-evaluation","slug":"claudemarketplaces-llm-evaluation","name":"llm-evaluation","type":"skill","source":"claudemarketplaces","status":"pending","summary":"This is a solid reference for measuring LLM output quality across automated metrics (BLEU, ROUGE, BERTScore), human evaluation rubrics, and LLM-as-judge patterns. You get working Python examples for scoring translations and summaries, comparing model outputs pairwise, and building custom metrics like groundedness checks. The automated metrics are fast but o...","profileUrl":"https://agentstaffer.com/skills/claudemarketplaces-llm-evaluation","fitScore":33,"roleFamily":"Coding & Development","reviewSignal":"Pending","urls":{"profileJson":"https://agentstaffer.com/skills/claudemarketplaces-llm-evaluation/profile.json","installJson":"https://agentstaffer.com/skills/claudemarketplaces-llm-evaluation/install.json","trustJson":"https://agentstaffer.com/skills/claudemarketplaces-llm-evaluation/trust.json","trialJson":"https://agentstaffer.com/skills/claudemarketplaces-llm-evaluation/trial.json"},"queueRank":13,"recommendation":{"label":"Needs review","score":33},"trialPriority":53,"trial":{"title":"Run a small code review trial","input":"Give the agent one low-risk issue or one small pull request and ask for defects, tests, and a patch plan.","expectedOutput":"A concise review with reproducible findings, changed-file scope, and test suggestions.","safetyNotes":["Run in a sandbox or throwaway workspace.","Use non-production data and temporary credentials.","Capture output, logs, install friction, and any permission requests."],"decisionCriteria":["Finds concrete issues","Avoids broad rewrites","Explains tradeoffs","Can be evaluated without secrets"],"runUrl":"https://agentstaffer.com/skills/claudemarketplaces-llm-evaluation#trial","runPath":"/skills/claudemarketplaces-llm-evaluation#trial","createEndpoint":"https://agentstaffer.com/api/actions/run-trial","createBody":{"listingId":"claudemarketplaces:skills-sickn33-antigravity-awesome-skills-llm-evaluation","reviewer":"Trial reviewer","notes":"Sandbox run using mock data."}},"launchPacket":{"schemaVersion":"agentstaffer.trial-launch-packet-lite.v1","requiredPermissions":{"schemaVersion":"agentstaffer.trial-requirements.v1","principle":"Use the smallest sandbox permission set needed for the first reversible trial.","requirements":[{"id":"sandbox","label":"Sandbox workspace","detail":"Run the first trial in a throwaway directory or isolated environment.","severity":"required","text":"Sandbox workspace: Run the first trial in a throwaway directory or isolated environment."},{"id":"license-review","label":"License review","detail":"License metadata is missing from the current public profile.","severity":"review","text":"License review: License metadata is missing from the current public profile."}],"humanReadable":["Sandbox workspace: Run the first trial in a throwaway directory or isolated environment.","License review: License metadata is missing from the current public profile."]},"preflightGates":[{"id":"profile","label":"Profile record","href":"https://agentstaffer.com/skills/claudemarketplaces-llm-evaluation/profile.json","passed":true},{"id":"install","label":"Setup guide","href":"https://agentstaffer.com/skills/claudemarketplaces-llm-evaluation/install.json","passed":true},{"id":"sandbox","label":"Use sandbox data","passed":true}],"promotionGates":[{"id":"install","label":"Install path reviewed","passed":true},{"id":"trust","label":"Trust evidence reviewed","passed":false},{"id":"trial","label":"Sandbox trial recorded","passed":false}],"stopConditions":["Install asks for production credentials or broad write access.","Source or package cannot be verified from profile evidence."],"createRequest":{"endpoint":"https://agentstaffer.com/api/actions/run-trial","method":"POST","body":{"listingId":"claudemarketplaces:skills-sickn33-antigravity-awesome-skills-llm-evaluation","reviewer":"Trial reviewer","notes":"Sandbox run using mock data."}}},"safety":{"trustTier":"Needs review","installReadiness":"Source guide","topRisk":{"level":"medium","label":"No community evaluation","detail":"AgentStaffer users have not submitted a review yet."},"requirements":[{"id":"sandbox","label":"Sandbox workspace","detail":"Run the first trial in a throwaway directory or isolated environment.","severity":"required"},{"id":"license-review","label":"License review","detail":"License metadata is missing from the current public profile.","severity":"review"}],"permissionSummary":{"schemaVersion":"agentstaffer.trial-requirements.v1","principle":"Use the smallest sandbox permission set needed for the first reversible trial.","requirements":[{"id":"sandbox","label":"Sandbox workspace","detail":"Run the first trial in a throwaway directory or isolated environment.","severity":"required","text":"Sandbox workspace: Run the first trial in a throwaway directory or isolated environment."},{"id":"license-review","label":"License review","detail":"License metadata is missing from the current public profile.","severity":"review","text":"License review: License metadata is missing from the current public profile."}],"humanReadable":["Sandbox workspace: Run the first trial in a throwaway directory or isolated environment.","License review: License metadata is missing from the current public profile."]}}},{"id":"github:msoedov/agentic_security","slug":"github-msoedov-agentic-security","name":"msoedov/agentic_security","type":"model","source":"github","status":"pending","summary":"Agentic LLM Vulnerability Scanner / AI red teaming kit 🧪","profileUrl":"https://agentstaffer.com/agents/github-msoedov-agentic-security","fitScore":37,"roleFamily":"Workflow Automation","reviewSignal":"Pending","urls":{"profileJson":"https://agentstaffer.com/agents/github-msoedov-agentic-security/profile.json","installJson":"https://agentstaffer.com/agents/github-msoedov-agentic-security/install.json","trustJson":"https://agentstaffer.com/agents/github-msoedov-agentic-security/trust.json","trialJson":"https://agentstaffer.com/agents/github-msoedov-agentic-security/trial.json"},"queueRank":14,"recommendation":{"label":"Needs review","score":37},"trialPriority":57,"trial":{"title":"Run a dry-run workflow trial","input":"Connect the workflow to mock inputs and ask it to produce a trace without touching production systems.","expectedOutput":"A step-by-step run log with inputs, outputs, failures, and rollback notes.","safetyNotes":["Run in a sandbox or throwaway workspace.","Use non-production data and temporary credentials.","Capture output, logs, install friction, and any permission requests."],"decisionCriteria":["Shows observable steps","Handles failure paths","Documents required credentials","Can run without side effects"],"runUrl":"https://agentstaffer.com/agents/github-msoedov-agentic-security#trial","runPath":"/agents/github-msoedov-agentic-security#trial","createEndpoint":"https://agentstaffer.com/api/actions/run-trial","createBody":{"listingId":"github:msoedov/agentic_security","reviewer":"Trial reviewer","notes":"Sandbox run using mock data."}},"launchPacket":{"schemaVersion":"agentstaffer.trial-launch-packet-lite.v1","requiredPermissions":{"schemaVersion":"agentstaffer.trial-requirements.v1","principle":"Use the smallest sandbox permission set needed for the first reversible trial.","requirements":[{"id":"sandbox","label":"Sandbox workspace","detail":"Run the first trial in a throwaway directory or isolated environment.","severity":"required","text":"Sandbox workspace: Run the first trial in a throwaway directory or isolated environment."}],"humanReadable":["Sandbox workspace: Run the first trial in a throwaway directory or isolated environment."]},"preflightGates":[{"id":"profile","label":"Profile record","href":"https://agentstaffer.com/agents/github-msoedov-agentic-security/profile.json","passed":true},{"id":"install","label":"Setup guide","href":"https://agentstaffer.com/agents/github-msoedov-agentic-security/install.json","passed":true},{"id":"sandbox","label":"Use sandbox data","passed":true}],"promotionGates":[{"id":"install","label":"Install path reviewed","passed":true},{"id":"trust","label":"Trust evidence reviewed","passed":false},{"id":"trial","label":"Sandbox trial recorded","passed":false}],"stopConditions":["Install asks for production credentials or broad write access.","Source or package cannot be verified from profile evidence."],"createRequest":{"endpoint":"https://agentstaffer.com/api/actions/run-trial","method":"POST","body":{"listingId":"github:msoedov/agentic_security","reviewer":"Trial reviewer","notes":"Sandbox run using mock data."}}},"safety":{"trustTier":"Needs review","installReadiness":"Source guide","topRisk":{"level":"medium","label":"No community evaluation","detail":"AgentStaffer users have not submitted a review yet."},"requirements":[{"id":"sandbox","label":"Sandbox workspace","detail":"Run the first trial in a throwaway directory or isolated environment.","severity":"required"}],"permissionSummary":{"schemaVersion":"agentstaffer.trial-requirements.v1","principle":"Use the smallest sandbox permission set needed for the first reversible trial.","requirements":[{"id":"sandbox","label":"Sandbox workspace","detail":"Run the first trial in a throwaway directory or isolated environment.","severity":"required","text":"Sandbox workspace: Run the first trial in a throwaway directory or isolated environment."}],"humanReadable":["Sandbox workspace: Run the first trial in a throwaway directory or isolated environment."]}}},{"id":"smithery:skills-microck-llm-evaluation","slug":"smithery-llm-evaluation","name":"LLM Evaluation","type":"skill","source":"smithery","status":"pending","summary":"LLM Evaluation is a Smithery skill listing by Microck.","profileUrl":"https://agentstaffer.com/skills/smithery-llm-evaluation","fitScore":24,"roleFamily":"Evaluation & Safety","reviewSignal":"Pending","urls":{"profileJson":"https://agentstaffer.com/skills/smithery-llm-evaluation/profile.json","installJson":"https://agentstaffer.com/skills/smithery-llm-evaluation/install.json","trustJson":"https://agentstaffer.com/skills/smithery-llm-evaluation/trust.json","trialJson":"https://agentstaffer.com/skills/smithery-llm-evaluation/trial.json"},"queueRank":15,"recommendation":{"label":"Needs review","score":24},"trialPriority":44,"trial":{"title":"Run a sandbox capability trial","input":"Ask the agent to complete one narrow task that reflects your intended use, using public or mock data only.","expectedOutput":"A useful result, clear assumptions, visible limitations, and enough evidence for a human evaluation.","safetyNotes":["Run in a sandbox or throwaway workspace.","Use non-production data and temporary credentials.","Capture output, logs, install friction, and any permission requests."],"decisionCriteria":["Completes the narrow task","Explains limits","Requires minimal privileges","Produces reviewable evidence"],"runUrl":"https://agentstaffer.com/skills/smithery-llm-evaluation#trial","runPath":"/skills/smithery-llm-evaluation#trial","createEndpoint":"https://agentstaffer.com/api/actions/run-trial","createBody":{"listingId":"smithery:skills-microck-llm-evaluation","reviewer":"Trial reviewer","notes":"Sandbox run using mock data."}},"launchPacket":{"schemaVersion":"agentstaffer.trial-launch-packet-lite.v1","requiredPermissions":{"schemaVersion":"agentstaffer.trial-requirements.v1","principle":"Use the smallest sandbox permission set needed for the first reversible trial.","requirements":[{"id":"sandbox","label":"Sandbox workspace","detail":"Run the first trial in a throwaway directory or isolated environment.","severity":"required","text":"Sandbox workspace: Run the first trial in a throwaway directory or isolated environment."},{"id":"license-review","label":"License review","detail":"License metadata is missing from the current public profile.","severity":"review","text":"License review: License metadata is missing from the current public profile."}],"humanReadable":["Sandbox workspace: Run the first trial in a throwaway directory or isolated environment.","License review: License metadata is missing from the current public profile."]},"preflightGates":[{"id":"profile","label":"Profile record","href":"https://agentstaffer.com/skills/smithery-llm-evaluation/profile.json","passed":true},{"id":"install","label":"Setup guide","href":"https://agentstaffer.com/skills/smithery-llm-evaluation/install.json","passed":true},{"id":"sandbox","label":"Use sandbox data","passed":true}],"promotionGates":[{"id":"install","label":"Install path reviewed","passed":true},{"id":"trust","label":"Trust evidence reviewed","passed":false},{"id":"trial","label":"Sandbox trial recorded","passed":false}],"stopConditions":["Install asks for production credentials or broad write access.","Source or package cannot be verified from profile evidence."],"createRequest":{"endpoint":"https://agentstaffer.com/api/actions/run-trial","method":"POST","body":{"listingId":"smithery:skills-microck-llm-evaluation","reviewer":"Trial reviewer","notes":"Sandbox run using mock data."}}},"safety":{"trustTier":"Needs review","installReadiness":"Source guide","topRisk":{"level":"medium","label":"No community evaluation","detail":"AgentStaffer users have not submitted a review yet."},"requirements":[{"id":"sandbox","label":"Sandbox workspace","detail":"Run the first trial in a throwaway directory or isolated environment.","severity":"required"},{"id":"license-review","label":"License review","detail":"License metadata is missing from the current public profile.","severity":"review"}],"permissionSummary":{"schemaVersion":"agentstaffer.trial-requirements.v1","principle":"Use the smallest sandbox permission set needed for the first reversible trial.","requirements":[{"id":"sandbox","label":"Sandbox workspace","detail":"Run the first trial in a throwaway directory or isolated environment.","severity":"required","text":"Sandbox workspace: Run the first trial in a throwaway directory or isolated environment."},{"id":"license-review","label":"License review","detail":"License metadata is missing from the current public profile.","severity":"review","text":"License review: License metadata is missing from the current public profile."}],"humanReadable":["Sandbox workspace: Run the first trial in a throwaway directory or isolated environment.","License review: License metadata is missing from the current public profile."]}}},{"id":"smithery:skills-yonatangross-llm-evaluation","slug":"smithery-llm-evaluation","name":"LLM Evaluation","type":"skill","source":"smithery","status":"pending","summary":"LLM Evaluation is a Smithery skill listing by yonatangross.","profileUrl":"https://agentstaffer.com/skills/smithery-llm-evaluation","fitScore":24,"roleFamily":"Evaluation & Safety","reviewSignal":"Pending","urls":{"profileJson":"https://agentstaffer.com/skills/smithery-llm-evaluation/profile.json","installJson":"https://agentstaffer.com/skills/smithery-llm-evaluation/install.json","trustJson":"https://agentstaffer.com/skills/smithery-llm-evaluation/trust.json","trialJson":"https://agentstaffer.com/skills/smithery-llm-evaluation/trial.json"},"queueRank":16,"recommendation":{"label":"Needs review","score":24},"trialPriority":44,"trial":{"title":"Run a sandbox capability trial","input":"Ask the agent to complete one narrow task that reflects your intended use, using public or mock data only.","expectedOutput":"A useful result, clear assumptions, visible limitations, and enough evidence for a human evaluation.","safetyNotes":["Run in a sandbox or throwaway workspace.","Use non-production data and temporary credentials.","Capture output, logs, install friction, and any permission requests."],"decisionCriteria":["Completes the narrow task","Explains limits","Requires minimal privileges","Produces reviewable evidence"],"runUrl":"https://agentstaffer.com/skills/smithery-llm-evaluation#trial","runPath":"/skills/smithery-llm-evaluation#trial","createEndpoint":"https://agentstaffer.com/api/actions/run-trial","createBody":{"listingId":"smithery:skills-yonatangross-llm-evaluation","reviewer":"Trial reviewer","notes":"Sandbox run using mock data."}},"launchPacket":{"schemaVersion":"agentstaffer.trial-launch-packet-lite.v1","requiredPermissions":{"schemaVersion":"agentstaffer.trial-requirements.v1","principle":"Use the smallest sandbox permission set needed for the first reversible trial.","requirements":[{"id":"sandbox","label":"Sandbox workspace","detail":"Run the first trial in a throwaway directory or isolated environment.","severity":"required","text":"Sandbox workspace: Run the first trial in a throwaway directory or isolated environment."},{"id":"license-review","label":"License review","detail":"License metadata is missing from the current public profile.","severity":"review","text":"License review: License metadata is missing from the current public profile."}],"humanReadable":["Sandbox workspace: Run the first trial in a throwaway directory or isolated environment.","License review: License metadata is missing from the current public profile."]},"preflightGates":[{"id":"profile","label":"Profile record","href":"https://agentstaffer.com/skills/smithery-llm-evaluation/profile.json","passed":true},{"id":"install","label":"Setup guide","href":"https://agentstaffer.com/skills/smithery-llm-evaluation/install.json","passed":true},{"id":"sandbox","label":"Use sandbox data","passed":true}],"promotionGates":[{"id":"install","label":"Install path reviewed","passed":true},{"id":"trust","label":"Trust evidence reviewed","passed":false},{"id":"trial","label":"Sandbox trial recorded","passed":false}],"stopConditions":["Install asks for production credentials or broad write access.","Source or package cannot be verified from profile evidence."],"createRequest":{"endpoint":"https://agentstaffer.com/api/actions/run-trial","method":"POST","body":{"listingId":"smithery:skills-yonatangross-llm-evaluation","reviewer":"Trial reviewer","notes":"Sandbox run using mock data."}}},"safety":{"trustTier":"Needs review","installReadiness":"Source guide","topRisk":{"level":"medium","label":"No community evaluation","detail":"AgentStaffer users have not submitted a review yet."},"requirements":[{"id":"sandbox","label":"Sandbox workspace","detail":"Run the first trial in a throwaway directory or isolated environment.","severity":"required"},{"id":"license-review","label":"License review","detail":"License metadata is missing from the current public profile.","severity":"review"}],"permissionSummary":{"schemaVersion":"agentstaffer.trial-requirements.v1","principle":"Use the smallest sandbox permission set needed for the first reversible trial.","requirements":[{"id":"sandbox","label":"Sandbox workspace","detail":"Run the first trial in a throwaway directory or isolated environment.","severity":"required","text":"Sandbox workspace: Run the first trial in a throwaway directory or isolated environment."},{"id":"license-review","label":"License review","detail":"License metadata is missing from the current public profile.","severity":"review","text":"License review: License metadata is missing from the current public profile."}],"humanReadable":["Sandbox workspace: Run the first trial in a throwaway directory or isolated environment.","License review: License metadata is missing from the current public profile."]}}},{"id":"github:palico-ai/palico-ai","slug":"github-palico-ai-palico-ai","name":"palico-ai/palico-ai","type":"model","source":"github","status":"pending","summary":"Build, Improve Performance, and Productionize your LLM Application with an Integrated Framework","profileUrl":"https://agentstaffer.com/agents/github-palico-ai-palico-ai","fitScore":31,"roleFamily":"Multi-agent Orchestration","reviewSignal":"Pending","urls":{"profileJson":"https://agentstaffer.com/agents/github-palico-ai-palico-ai/profile.json","installJson":"https://agentstaffer.com/agents/github-palico-ai-palico-ai/install.json","trustJson":"https://agentstaffer.com/agents/github-palico-ai-palico-ai/trust.json","trialJson":"https://agentstaffer.com/agents/github-palico-ai-palico-ai/trial.json"},"queueRank":17,"recommendation":{"label":"Needs review","score":31},"trialPriority":51,"trial":{"title":"Run a sandbox capability trial","input":"Ask the agent to complete one narrow task that reflects your intended use, using public or mock data only.","expectedOutput":"A useful result, clear assumptions, visible limitations, and enough evidence for a human evaluation.","safetyNotes":["Run in a sandbox or throwaway workspace.","Use non-production data and temporary credentials.","Capture output, logs, install friction, and any permission requests."],"decisionCriteria":["Completes the narrow task","Explains limits","Requires minimal privileges","Produces reviewable evidence"],"runUrl":"https://agentstaffer.com/agents/github-palico-ai-palico-ai#trial","runPath":"/agents/github-palico-ai-palico-ai#trial","createEndpoint":"https://agentstaffer.com/api/actions/run-trial","createBody":{"listingId":"github:palico-ai/palico-ai","reviewer":"Trial reviewer","notes":"Sandbox run using mock data."}},"launchPacket":{"schemaVersion":"agentstaffer.trial-launch-packet-lite.v1","requiredPermissions":{"schemaVersion":"agentstaffer.trial-requirements.v1","principle":"Use the smallest sandbox permission set needed for the first reversible trial.","requirements":[{"id":"sandbox","label":"Sandbox workspace","detail":"Run the first trial in a throwaway directory or isolated environment.","severity":"required","text":"Sandbox workspace: Run the first trial in a throwaway directory or isolated environment."},{"id":"container","label":"Container runtime","detail":"Docker or Kubernetes appears in metadata. Confirm the runtime before trial.","severity":"conditional","text":"Container runtime: Docker or Kubernetes appears in metadata. Confirm the runtime before trial."},{"id":"temporary-credentials","label":"Temporary credentials","detail":"Use throwaway API keys or scoped test credentials. Do not start with production secrets.","severity":"conditional","text":"Temporary credentials: Use throwaway API keys or scoped test credentials. Do not start with production secrets."}],"humanReadable":["Sandbox workspace: Run the first trial in a throwaway directory or isolated environment.","Container runtime: Docker or Kubernetes appears in metadata. Confirm the runtime before trial.","Temporary credentials: Use throwaway API keys or scoped test credentials. Do not start with production secrets."]},"preflightGates":[{"id":"profile","label":"Profile record","href":"https://agentstaffer.com/agents/github-palico-ai-palico-ai/profile.json","passed":true},{"id":"install","label":"Setup guide","href":"https://agentstaffer.com/agents/github-palico-ai-palico-ai/install.json","passed":true},{"id":"sandbox","label":"Use sandbox data","passed":true}],"promotionGates":[{"id":"install","label":"Install path reviewed","passed":true},{"id":"trust","label":"Trust evidence reviewed","passed":false},{"id":"trial","label":"Sandbox trial recorded","passed":false}],"stopConditions":["Install asks for production credentials or broad write access.","Source or package cannot be verified from profile evidence."],"createRequest":{"endpoint":"https://agentstaffer.com/api/actions/run-trial","method":"POST","body":{"listingId":"github:palico-ai/palico-ai","reviewer":"Trial reviewer","notes":"Sandbox run using mock data."}}},"safety":{"trustTier":"Needs review","installReadiness":"Source guide","topRisk":{"level":"medium","label":"No community evaluation","detail":"AgentStaffer users have not submitted a review yet."},"requirements":[{"id":"sandbox","label":"Sandbox workspace","detail":"Run the first trial in a throwaway directory or isolated environment.","severity":"required"},{"id":"container","label":"Container runtime","detail":"Docker or Kubernetes appears in metadata. Confirm the runtime before trial.","severity":"conditional"},{"id":"temporary-credentials","label":"Temporary credentials","detail":"Use throwaway API keys or scoped test credentials. Do not start with production secrets.","severity":"conditional"}],"permissionSummary":{"schemaVersion":"agentstaffer.trial-requirements.v1","principle":"Use the smallest sandbox permission set needed for the first reversible trial.","requirements":[{"id":"sandbox","label":"Sandbox workspace","detail":"Run the first trial in a throwaway directory or isolated environment.","severity":"required","text":"Sandbox workspace: Run the first trial in a throwaway directory or isolated environment."},{"id":"container","label":"Container runtime","detail":"Docker or Kubernetes appears in metadata. Confirm the runtime before trial.","severity":"conditional","text":"Container runtime: Docker or Kubernetes appears in metadata. Confirm the runtime before trial."},{"id":"temporary-credentials","label":"Temporary credentials","detail":"Use throwaway API keys or scoped test credentials. Do not start with production secrets.","severity":"conditional","text":"Temporary credentials: Use throwaway API keys or scoped test credentials. Do not start with production secrets."}],"humanReadable":["Sandbox workspace: Run the first trial in a throwaway directory or isolated environment.","Container runtime: Docker or Kubernetes appears in metadata. Confirm the runtime before trial.","Temporary credentials: Use throwaway API keys or scoped test credentials. Do not start with production secrets."]}}},{"id":"github:imeanai/webcanvas","slug":"github-imeanai-webcanvas","name":"iMeanAI/WebCanvas","type":"model","source":"github","status":"pending","summary":"All-in-one Web Agent framework for post-training. Start building with a few clicks!","profileUrl":"https://agentstaffer.com/agents/github-imeanai-webcanvas","fitScore":29,"roleFamily":"Coding & Development","reviewSignal":"Pending","urls":{"profileJson":"https://agentstaffer.com/agents/github-imeanai-webcanvas/profile.json","installJson":"https://agentstaffer.com/agents/github-imeanai-webcanvas/install.json","trustJson":"https://agentstaffer.com/agents/github-imeanai-webcanvas/trust.json","trialJson":"https://agentstaffer.com/agents/github-imeanai-webcanvas/trial.json"},"queueRank":18,"recommendation":{"label":"Needs review","score":29},"trialPriority":49,"trial":{"title":"Run a small code review trial","input":"Give the agent one low-risk issue or one small pull request and ask for defects, tests, and a patch plan.","expectedOutput":"A concise review with reproducible findings, changed-file scope, and test suggestions.","safetyNotes":["Run in a sandbox or throwaway workspace.","Use non-production data and temporary credentials.","Capture output, logs, install friction, and any permission requests."],"decisionCriteria":["Finds concrete issues","Avoids broad rewrites","Explains tradeoffs","Can be evaluated without secrets"],"runUrl":"https://agentstaffer.com/agents/github-imeanai-webcanvas#trial","runPath":"/agents/github-imeanai-webcanvas#trial","createEndpoint":"https://agentstaffer.com/api/actions/run-trial","createBody":{"listingId":"github:imeanai/webcanvas","reviewer":"Trial reviewer","notes":"Sandbox run using mock data."}},"launchPacket":{"schemaVersion":"agentstaffer.trial-launch-packet-lite.v1","requiredPermissions":{"schemaVersion":"agentstaffer.trial-requirements.v1","principle":"Use the smallest sandbox permission set needed for the first reversible trial.","requirements":[{"id":"sandbox","label":"Sandbox workspace","detail":"Run the first trial in a throwaway directory or isolated environment.","severity":"required","text":"Sandbox workspace: Run the first trial in a throwaway directory or isolated environment."},{"id":"terminal","label":"Local terminal","detail":"A CLI path is visible, but the exact command must be verified from source.","severity":"required","text":"Local terminal: A CLI path is visible, but the exact command must be verified from source."}],"humanReadable":["Sandbox workspace: Run the first trial in a throwaway directory or isolated environment.","Local terminal: A CLI path is visible, but the exact command must be verified from source."]},"preflightGates":[{"id":"profile","label":"Profile record","href":"https://agentstaffer.com/agents/github-imeanai-webcanvas/profile.json","passed":true},{"id":"install","label":"Setup guide","href":"https://agentstaffer.com/agents/github-imeanai-webcanvas/install.json","passed":true},{"id":"sandbox","label":"Use sandbox data","passed":true}],"promotionGates":[{"id":"install","label":"Install path reviewed","passed":true},{"id":"trust","label":"Trust evidence reviewed","passed":false},{"id":"trial","label":"Sandbox trial recorded","passed":false}],"stopConditions":["Install asks for production credentials or broad write access.","Source or package cannot be verified from profile evidence."],"createRequest":{"endpoint":"https://agentstaffer.com/api/actions/run-trial","method":"POST","body":{"listingId":"github:imeanai/webcanvas","reviewer":"Trial reviewer","notes":"Sandbox run using mock data."}}},"safety":{"trustTier":"Needs review","installReadiness":"CLI likely","topRisk":{"level":"medium","label":"No community evaluation","detail":"AgentStaffer users have not submitted a review yet."},"requirements":[{"id":"sandbox","label":"Sandbox workspace","detail":"Run the first trial in a throwaway directory or isolated environment.","severity":"required"},{"id":"terminal","label":"Local terminal","detail":"A CLI path is visible, but the exact command must be verified from source.","severity":"required"}],"permissionSummary":{"schemaVersion":"agentstaffer.trial-requirements.v1","principle":"Use the smallest sandbox permission set needed for the first reversible trial.","requirements":[{"id":"sandbox","label":"Sandbox workspace","detail":"Run the first trial in a throwaway directory or isolated environment.","severity":"required","text":"Sandbox workspace: Run the first trial in a throwaway directory or isolated environment."},{"id":"terminal","label":"Local terminal","detail":"A CLI path is visible, but the exact command must be verified from source.","severity":"required","text":"Local terminal: A CLI path is visible, but the exact command must be verified from source."}],"humanReadable":["Sandbox workspace: Run the first trial in a throwaway directory or isolated environment.","Local terminal: A CLI path is visible, but the exact command must be verified from source."]}}},{"id":"github:open-compass/gta","slug":"github-open-compass-gta","name":"open-compass/GTA","type":"model","source":"github","status":"pending","summary":"[NeurIPS 2024 D&B] GTA: A Benchmark for General Tool Agents & [arXiv 2026] GTA-2","profileUrl":"https://agentstaffer.com/agents/github-open-compass-gta","fitScore":21,"roleFamily":"Evaluation & Safety","reviewSignal":"Pending","urls":{"profileJson":"https://agentstaffer.com/agents/github-open-compass-gta/profile.json","installJson":"https://agentstaffer.com/agents/github-open-compass-gta/install.json","trustJson":"https://agentstaffer.com/agents/github-open-compass-gta/trust.json","trialJson":"https://agentstaffer.com/agents/github-open-compass-gta/trial.json"},"queueRank":19,"recommendation":{"label":"Needs review","score":21},"trialPriority":41,"trial":{"title":"Run a sandbox capability trial","input":"Ask the agent to complete one narrow task that reflects your intended use, using public or mock data only.","expectedOutput":"A useful result, clear assumptions, visible limitations, and enough evidence for a human evaluation.","safetyNotes":["Run in a sandbox or throwaway workspace.","Use non-production data and temporary credentials.","Capture output, logs, install friction, and any permission requests."],"decisionCriteria":["Completes the narrow task","Explains limits","Requires minimal privileges","Produces reviewable evidence"],"runUrl":"https://agentstaffer.com/agents/github-open-compass-gta#trial","runPath":"/agents/github-open-compass-gta#trial","createEndpoint":"https://agentstaffer.com/api/actions/run-trial","createBody":{"listingId":"github:open-compass/gta","reviewer":"Trial reviewer","notes":"Sandbox run using mock data."}},"launchPacket":{"schemaVersion":"agentstaffer.trial-launch-packet-lite.v1","requiredPermissions":{"schemaVersion":"agentstaffer.trial-requirements.v1","principle":"Use the smallest sandbox permission set needed for the first reversible trial.","requirements":[{"id":"sandbox","label":"Sandbox workspace","detail":"Run the first trial in a throwaway directory or isolated environment.","severity":"required","text":"Sandbox workspace: Run the first trial in a throwaway directory or isolated environment."}],"humanReadable":["Sandbox workspace: Run the first trial in a throwaway directory or isolated environment."]},"preflightGates":[{"id":"profile","label":"Profile record","href":"https://agentstaffer.com/agents/github-open-compass-gta/profile.json","passed":true},{"id":"install","label":"Setup guide","href":"https://agentstaffer.com/agents/github-open-compass-gta/install.json","passed":true},{"id":"sandbox","label":"Use sandbox data","passed":true}],"promotionGates":[{"id":"install","label":"Install path reviewed","passed":true},{"id":"trust","label":"Trust evidence reviewed","passed":false},{"id":"trial","label":"Sandbox trial recorded","passed":false}],"stopConditions":["Install asks for production credentials or broad write access.","Source or package cannot be verified from profile evidence."],"createRequest":{"endpoint":"https://agentstaffer.com/api/actions/run-trial","method":"POST","body":{"listingId":"github:open-compass/gta","reviewer":"Trial reviewer","notes":"Sandbox run using mock data."}}},"safety":{"trustTier":"Needs review","installReadiness":"Source guide","topRisk":{"level":"medium","label":"No community evaluation","detail":"AgentStaffer users have not submitted a review yet."},"requirements":[{"id":"sandbox","label":"Sandbox workspace","detail":"Run the first trial in a throwaway directory or isolated environment.","severity":"required"}],"permissionSummary":{"schemaVersion":"agentstaffer.trial-requirements.v1","principle":"Use the smallest sandbox permission set needed for the first reversible trial.","requirements":[{"id":"sandbox","label":"Sandbox workspace","detail":"Run the first trial in a throwaway directory or isolated environment.","severity":"required","text":"Sandbox workspace: Run the first trial in a throwaway directory or isolated environment."}],"humanReadable":["Sandbox workspace: Run the first trial in a throwaway directory or isolated environment."]}}},{"id":"github:valuebyte-ai/awesome-llm-in-social-science","slug":"github-valuebyte-ai-awesome-llm-in-social-science","name":"ValueByte-AI/Awesome-LLM-in-Social-Science","type":"model","source":"github","status":"pending","summary":"Awesome papers involving LLMs in Social Science.","profileUrl":"https://agentstaffer.com/agents/github-valuebyte-ai-awesome-llm-in-social-science","fitScore":20,"roleFamily":"Evaluation & Safety","reviewSignal":"Pending","urls":{"profileJson":"https://agentstaffer.com/agents/github-valuebyte-ai-awesome-llm-in-social-science/profile.json","installJson":"https://agentstaffer.com/agents/github-valuebyte-ai-awesome-llm-in-social-science/install.json","trustJson":"https://agentstaffer.com/agents/github-valuebyte-ai-awesome-llm-in-social-science/trust.json","trialJson":"https://agentstaffer.com/agents/github-valuebyte-ai-awesome-llm-in-social-science/trial.json"},"queueRank":20,"recommendation":{"label":"Needs review","score":20},"trialPriority":40,"trial":{"title":"Run a sandbox capability trial","input":"Ask the agent to complete one narrow task that reflects your intended use, using public or mock data only.","expectedOutput":"A useful result, clear assumptions, visible limitations, and enough evidence for a human evaluation.","safetyNotes":["Run in a sandbox or throwaway workspace.","Use non-production data and temporary credentials.","Capture output, logs, install friction, and any permission requests."],"decisionCriteria":["Completes the narrow task","Explains limits","Requires minimal privileges","Produces reviewable evidence"],"runUrl":"https://agentstaffer.com/agents/github-valuebyte-ai-awesome-llm-in-social-science#trial","runPath":"/agents/github-valuebyte-ai-awesome-llm-in-social-science#trial","createEndpoint":"https://agentstaffer.com/api/actions/run-trial","createBody":{"listingId":"github:valuebyte-ai/awesome-llm-in-social-science","reviewer":"Trial reviewer","notes":"Sandbox run using mock data."}},"launchPacket":{"schemaVersion":"agentstaffer.trial-launch-packet-lite.v1","requiredPermissions":{"schemaVersion":"agentstaffer.trial-requirements.v1","principle":"Use the smallest sandbox permission set needed for the first reversible trial.","requirements":[{"id":"sandbox","label":"Sandbox workspace","detail":"Run the first trial in a throwaway directory or isolated environment.","severity":"required","text":"Sandbox workspace: Run the first trial in a throwaway directory or isolated environment."}],"humanReadable":["Sandbox workspace: Run the first trial in a throwaway directory or isolated environment."]},"preflightGates":[{"id":"profile","label":"Profile record","href":"https://agentstaffer.com/agents/github-valuebyte-ai-awesome-llm-in-social-science/profile.json","passed":true},{"id":"install","label":"Setup guide","href":"https://agentstaffer.com/agents/github-valuebyte-ai-awesome-llm-in-social-science/install.json","passed":true},{"id":"sandbox","label":"Use sandbox data","passed":true}],"promotionGates":[{"id":"install","label":"Install path reviewed","passed":true},{"id":"trust","label":"Trust evidence reviewed","passed":false},{"id":"trial","label":"Sandbox trial recorded","passed":false}],"stopConditions":["Install asks for production credentials or broad write access.","Source or package cannot be verified from profile evidence."],"createRequest":{"endpoint":"https://agentstaffer.com/api/actions/run-trial","method":"POST","body":{"listingId":"github:valuebyte-ai/awesome-llm-in-social-science","reviewer":"Trial reviewer","notes":"Sandbox run using mock data."}}},"safety":{"trustTier":"Needs review","installReadiness":"Source guide","topRisk":{"level":"medium","label":"No community evaluation","detail":"AgentStaffer users have not submitted a review yet."},"requirements":[{"id":"sandbox","label":"Sandbox workspace","detail":"Run the first trial in a throwaway directory or isolated environment.","severity":"required"}],"permissionSummary":{"schemaVersion":"agentstaffer.trial-requirements.v1","principle":"Use the smallest sandbox permission set needed for the first reversible trial.","requirements":[{"id":"sandbox","label":"Sandbox workspace","detail":"Run the first trial in a throwaway directory or isolated environment.","severity":"required","text":"Sandbox workspace: Run the first trial in a throwaway directory or isolated environment."}],"humanReadable":["Sandbox workspace: Run the first trial in a throwaway directory or isolated environment."]}}},{"id":"github:fuxiailab/civagent","slug":"github-fuxiailab-civagent","name":"fuxiAIlab/CivAgent","type":"model","source":"github","status":"pending","summary":"CivAgent is an LLM-based Human-like Agent acting as a Digital Player within the Strategy Game Unciv.","profileUrl":"https://agentstaffer.com/agents/github-fuxiailab-civagent","fitScore":19,"roleFamily":"Evaluation & Safety","reviewSignal":"Pending","urls":{"profileJson":"https://agentstaffer.com/agents/github-fuxiailab-civagent/profile.json","installJson":"https://agentstaffer.com/agents/github-fuxiailab-civagent/install.json","trustJson":"https://agentstaffer.com/agents/github-fuxiailab-civagent/trust.json","trialJson":"https://agentstaffer.com/agents/github-fuxiailab-civagent/trial.json"},"queueRank":21,"recommendation":{"label":"Needs review","score":19},"trialPriority":39,"trial":{"title":"Run a sandbox capability trial","input":"Ask the agent to complete one narrow task that reflects your intended use, using public or mock data only.","expectedOutput":"A useful result, clear assumptions, visible limitations, and enough evidence for a human evaluation.","safetyNotes":["Run in a sandbox or throwaway workspace.","Use non-production data and temporary credentials.","Capture output, logs, install friction, and any permission requests."],"decisionCriteria":["Completes the narrow task","Explains limits","Requires minimal privileges","Produces reviewable evidence"],"runUrl":"https://agentstaffer.com/agents/github-fuxiailab-civagent#trial","runPath":"/agents/github-fuxiailab-civagent#trial","createEndpoint":"https://agentstaffer.com/api/actions/run-trial","createBody":{"listingId":"github:fuxiailab/civagent","reviewer":"Trial reviewer","notes":"Sandbox run using mock data."}},"launchPacket":{"schemaVersion":"agentstaffer.trial-launch-packet-lite.v1","requiredPermissions":{"schemaVersion":"agentstaffer.trial-requirements.v1","principle":"Use the smallest sandbox permission set needed for the first reversible trial.","requirements":[{"id":"sandbox","label":"Sandbox workspace","detail":"Run the first trial in a throwaway directory or isolated environment.","severity":"required","text":"Sandbox workspace: Run the first trial in a throwaway directory or isolated environment."}],"humanReadable":["Sandbox workspace: Run the first trial in a throwaway directory or isolated environment."]},"preflightGates":[{"id":"profile","label":"Profile record","href":"https://agentstaffer.com/agents/github-fuxiailab-civagent/profile.json","passed":true},{"id":"install","label":"Setup guide","href":"https://agentstaffer.com/agents/github-fuxiailab-civagent/install.json","passed":true},{"id":"sandbox","label":"Use sandbox data","passed":true}],"promotionGates":[{"id":"install","label":"Install path reviewed","passed":true},{"id":"trust","label":"Trust evidence reviewed","passed":false},{"id":"trial","label":"Sandbox trial recorded","passed":false}],"stopConditions":["Install asks for production credentials or broad write access.","Source or package cannot be verified from profile evidence."],"createRequest":{"endpoint":"https://agentstaffer.com/api/actions/run-trial","method":"POST","body":{"listingId":"github:fuxiailab/civagent","reviewer":"Trial reviewer","notes":"Sandbox run using mock data."}}},"safety":{"trustTier":"Needs review","installReadiness":"Source guide","topRisk":{"level":"medium","label":"No community evaluation","detail":"AgentStaffer users have not submitted a review yet."},"requirements":[{"id":"sandbox","label":"Sandbox workspace","detail":"Run the first trial in a throwaway directory or isolated environment.","severity":"required"}],"permissionSummary":{"schemaVersion":"agentstaffer.trial-requirements.v1","principle":"Use the smallest sandbox permission set needed for the first reversible trial.","requirements":[{"id":"sandbox","label":"Sandbox workspace","detail":"Run the first trial in a throwaway directory or isolated environment.","severity":"required","text":"Sandbox workspace: Run the first trial in a throwaway directory or isolated environment."}],"humanReadable":["Sandbox workspace: Run the first trial in a throwaway directory or isolated environment."]}}}],"recentRuns":[],"pagination":{"page":1,"pageSize":24,"totalItems":21,"totalPages":1,"nextPageUrl":null,"prevPageUrl":null},"filters":{"q":"llm-evaluation","type":null,"role":null,"protocol":null,"trust":null,"install":null,"source":null,"status":null,"sort":"trial_readiness"},"availableFilters":{"roles":[{"id":"coding-development","label":"Coding & Development"},{"id":"workflow-automation","label":"Workflow Automation"},{"id":"multi-agent-orchestration","label":"Multi-agent Orchestration"},{"id":"mcp-tool-integration","label":"MCP & Tool Integration"},{"id":"research-knowledge","label":"Research & Knowledge"},{"id":"computer-use-desktop","label":"Computer Use & Desktop"},{"id":"business-operations","label":"Business Operations"},{"id":"data-integrations","label":"Data & Integrations"},{"id":"evaluation-safety","label":"Evaluation & Safety"},{"id":"model-inference","label":"Models & Inference"},{"id":"general-agent","label":"General Agent"}],"protocols":[{"value":"mcp","label":"MCP (9)"},{"value":"cli","label":"CLI (6)"},{"value":"http","label":"HTTP (5)"},{"value":"sdk","label":"SDK (2)"}],"trust":[{"value":"needs-review","label":"Needs review (20)"},{"value":"verified-source","label":"Verified source (1)"}],"install":[{"value":"source-guide","label":"Source guide (11)"},{"value":"mcp-ready","label":"MCP ready (9)"},{"value":"cli-likely","label":"CLI likely (1)"}],"sources":[{"value":"github","label":"github (8)"},{"value":"playbooks","label":"playbooks (3)"},{"value":"remoteopenclaw","label":"remoteopenclaw (2)"},{"value":"smithery","label":"smithery (2)"},{"value":"agensi","label":"agensi (1)"},{"value":"claudemarketplaces","label":"claudemarketplaces (1)"},{"value":"enterprisedna","label":"enterprisedna (1)"},{"value":"explainx","label":"explainx (1)"},{"value":"mcpservers","label":"mcpservers (1)"},{"value":"npm","label":"npm (1)"}],"types":[{"value":"skill","label":"skill (11)"},{"value":"model","label":"model (7)"},{"value":"workflow","label":"workflow (2)"},{"value":"framework","label":"framework (1)"}],"sortModes":["relevance","trust","updated","popularity","install_readiness","trial_readiness"]},"launchProtocol":{"schemaVersion":"agentstaffer.trial-launch-packet.v1","location":"trialQueue[].launchPacket","permissionsLocation":"trialQueue[].safety.permissionSummary","rule":"Use launchPacket.preflightGates and launchPacket.stopConditions before calling launchPacket.createRequest.","promotionRule":"Do not promote or hire until launchPacket.promotionGates are clear or explicitly accepted by a human reviewer."},"visibleRetrievalSpine":[{"step":"01","label":"llms.txt","href":"/llms.txt","note":"Site policy and safety context"},{"step":"02","label":"trials.json","href":"/trials.json?q=llm-evaluation&sort=trial_readiness","note":"Action queue and page contract"},{"step":"03","label":"profile HTML","href":"/mcp/npm-confused-ai","note":"Human decision summary"},{"step":"04","label":"profile package","href":"/mcp/npm-confused-ai/profile.json","note":"Fetch all 9 required finalist resources","requiredResources":["profile.json","score.json","evidence.json","trust.json","trial.json","install.json","history.json","similar.json","dossier.md"]}],"profileResourceContract":{"schemaVersion":"agentstaffer.profile-resource-contract.v1","requiredResources":["profile.json","score.json","evidence.json","trust.json","trial.json","install.json","history.json","similar.json","dossier.md"],"recommendedFetchOrder":["profile.json","score.json","evidence.json","trust.json","trial.json","install.json","history.json","similar.json","dossier.md"]},"retrievalHints":["Use trialQueue[].trial before recommending hire or production installation.","Use trialQueue[].safety.permissionSummary.humanReadable when explaining first-run permissions to a human.","Use trialQueue[].launchPacket as the copyable trial launch contract for humans and AI agents.","POST to /api/actions/run-trial with trialQueue[].trial.createBody to create a public trial result.","Fetch trialQueue[].urls.profileJson (profile.json) and trialQueue[].urls.installJson (install.json) for the full dossier and install payloads.","Use recentRuns[] as auditable evidence of completed or blocked sandbox evaluations."],"links":{"trials":"https://agentstaffer.com/trials?q=llm-evaluation&sort=trial_readiness","trialsJson":"https://agentstaffer.com/trials.json?q=llm-evaluation&sort=trial_readiness","compare":"https://agentstaffer.com/compare","promotions":"https://agentstaffer.com/promotions","jobMatcher":"https://agentstaffer.com/job-matcher","openApi":"https://agentstaffer.com/openapi.json","llmsTxt":"https://agentstaffer.com/llms.txt","geoJson":"https://agentstaffer.com/geo.json","agentVisibilityJson":"https://agentstaffer.com/agent-visibility.json","aiCrawlMonitorJson":"https://agentstaffer.com/ai-crawl-monitor.json"}}