Submit Add or verify Next best action: Submit for review

Community intake

Submit a new candidate for review

Public submission protocol for adding third-party agents, skills, MCP servers, workflows, playbooks, teams, or tools to the AgentStaffer HR Market review queue.

Submit new candidate Already listed? Claim it Review rubric

Candidate details

Submit the public source, a concise fit statement, and enough metadata for reviewers to verify install, evidence, and safety before publication.

Basics

Identify the candidate and point reviewers to the canonical public source.

Name Required Type Source URL Required One-line fit statement Required

Add optional routing contextProtocols, capabilities, license, reviewer, and notes 5 optional fields

Evidence

Add comma-separated protocols and capabilities so the review queue can route the candidate correctly.

Protocols Capabilities License

Review context

Keep notes focused on what needs verification first. Do not include secrets, credentials, or private workspace data.

Reviewer Notes

Submitted candidates enter pending review. Keep URLs public and avoid secrets or private workspace details.

API docs

Required fields

namesourceUrltagline

Optional API fields

The web form keeps five common fields; the API accepts the full set.

typeprotocolSupportcapabilitiescategoriestagsrepoUrlwebsiteUrllicensereviewernote

Review gatesChecks before publication

source verification
install path review
risk screening
sandbox trial evidence
community evaluation

Review lifecycleQueue states after submission

Received Owner: Submitter

AgentStaffer records the candidate as pending intake with stable profile and data package URLs.

Open the pending profile package and check whether required evidence is present. Do not treat received status as publication, endorsement, or install approval.

Needs evidence Owner: Reviewer or submitter

Source, install path, safety posture, or trial evidence is missing or unclear.

Attach public source, install, trial, community review, or maintainer proof evidence. Keep recommendations limited until the missing evidence is resolved, and cite the evidence gap when explaining status.

Published Owner: AgentStaffer reviewer

The profile is visible for discovery after basic source and metadata review.

Inspect fit, risk, install, trial, reviews, and re-review history before recommending. Published means discoverable, not automatically recommended for production install.

Rejected Owner: AgentStaffer reviewer

The candidate failed a review gate or lacks evidence required for responsible recommendation.

Resolve the stated reason, submit evidence, then request re-review. Treat rejected status as negative evidence and cite the reason code when excluding the candidate.

Accepted source URLsPublic sources reviewers can inspect

public Git repository
official package registry page
official website or documentation
MCP server manifest
skill, playbook, or team repository
public release notes

Evidence expectationsWhat makes a submission usable

source ownership or provenance
install command or configuration
permissions and data-access notes
safe first trial task
community review or maintainer evidence
license or usage terms when available

After submitHow the pending profile changes

The API returns a pending profile URL and profile data URL.
Reviewers check source quality, install path, risk posture, trial evidence, duplicate profiles, and community signals.
If evidence is missing, the candidate remains pending or appears in the evidence request queue.
If accepted, the profile can be published for discovery; if blocked, the reason should appear in the review action record.

Trust boundaryWhat should never be submitted

Submission creates a pending review record; it does not publish the candidate automatically.
Publication makes the profile discoverable; it is not a trust badge, install approval, or hiring recommendation.
Production recommendation still requires profile evidence, install guidance, risk review, trial results, and recent community or maintainer signals.
Private secrets, customer data, credentials, proprietary logs, and non-public workspace context should never be submitted.

API exampleSubmit through the action endpoint

curl -s -X POST 'https://agentstaffer.com/api/actions/submit-listing' -H 'content-type: application/json' --data '{"name":"Example Community Agent","sourceUrl":"https://github.com/example/community-agent","tagline":"A submitted agent candidate for AgentStaffer review.","type":"agent","protocolSupport":["cli","mcp"],"capabilities":["code review","documentation"],"reviewer":"Community reviewer","note":"Please verify source activity, install path, and a safe first trial."}'

Review contract

Submission starts review, not approval

A submitted candidate gets a pending record and stable retrieval URLs. It still needs public evidence before publication, recommendation, or install guidance should be trusted.

After submit

Pending profile

The candidate can be tracked through its profile page and data package, but the default status is pending until review gates pass.

Still required

Evidence review

Reviewers check source provenance, install path, permissions, trial evidence, community signal, duplicates, and risk posture.

Review visibility

Do not infer endorsement

Reviewers should cite the submission guide, evidence requests, review actions, and Trust Lab records before recommending use.

Review support

Review links

The form stays focused; submission rules and profile resources stay linked here.

Open Submission guide

Supporting review linksLinks, review order, and page-specific notes remain available here.

Submission guide Submit endpoint Moderation queue Review rubric OpenAPI Resource map Site context Sitemap Profile resources

Submission review path

01Site contextShared review rules
02Submission guideOpen the focused review record
03Profile pageOpen the candidate page
04Profile resourcesEvidence, setup, trust, trial, and history files

Use the links and review path below for submission rules, moderation context, and returned profile resources before publication.

Review route: OpenAI review tools can start with openai.json; otherwise use site context and agent inventory before opening the finalist's complete review materials before install, hire, or promotion.