Market / Sources Map source evidence Next best action: Review ecosystem evidence, then fetch finalist profile resources before recommending

Source Directory

Evaluate agent ecosystems before individual candidates

A practical overview of where candidates come from, how each ecosystem looks on evidence and risk, and which candidates deserve deeper review.

Browse market Open Trust Lab

Coverage boundary

Currently segmented: Playbooks, npm, GitHub, Awesomeskill, Explainx, Mcpservers

Current public source segmentation covers 6 source groups. Use each source row as an evidence starting point, not a final safety signal.

Some ecosystems may still be represented through linked source profiles before they become separate rows.
Source popularity is only a discovery signal, not a safety or maintenance guarantee.
Open the complete finalist review set before recommending setup, hire, or credential access.

playbooks Playbooks

Treat this source as exploration only until profile trust, install, evidence, and trial resources are fetched.

13listings

Quality54 / 100

Evidence0%

Install ready13

Needs review13

ai-agents skill by vasilyu1983/ai-agents-publicNeeds review · MCP ready langchain-docs skill by jokejason/local-context7Needs review · MCP ready langgraph-multiagent skill by bjornmelin/dev-skillsNeeds review · MCP ready

Source reviewQuality signals, common risks, install caveats, and trust mix for Playbooks.

Needs source review 54 / 100 source quality

This source has weaker standardized provenance. Treat it as a discovery route until profile-level evidence is fetched.

Source metrics

Source quality: 54 / 100
Risk rate: 100%
Evidence coverage: 0%
Average fit: 65 / 100
Install ready: 13
MCP ready: 13
Skills: 13
Needs review: 13
Risk flagged: 13

Common risks

unknown maintainer status
incomplete install metadata
missing community review
unclear update cadence

Install caveats

Fetch trust.json and evidence.json before install.
Use manual review for the first candidate.
Do not infer safety from source label alone.

ai-agents skill by vasilyu1983/ai-agents-public

langchain-docs skill by jokejason/local-context7

langgraph-multiagent skill by bjornmelin/dev-skills

13Needs review

Candidate set MCP-ready list Skill packages Search

npm npm

Treat this source as exploration only until profile trust, install, evidence, and trial resources are fetched.

6listings

Quality53 / 100

Evidence0%

Install ready6

Needs review6

iflow-mcp-sixhq-overtureNeeds review · MCP ready overture-mcpNeeds review · MCP ready awaithumansNeeds review · MCP ready

Source reviewQuality signals, common risks, install caveats, and trust mix for npm.

Needs source review 53 / 100 source quality

npm metadata helps package discovery and quick install, but package ownership, dependency scripts, and source linkage need separate verification.

Source metrics

Source quality: 53 / 100
Risk rate: 100%
Evidence coverage: 0%
Average fit: 60 / 100
Install ready: 6
MCP ready: 6
Skills: 3
Needs review: 6
Risk flagged: 6

Common risks

package takeover
postinstall scripts
transitive dependency drift
package name confusion

Install caveats

Verify repository link and package owner.
Inspect lifecycle scripts before install.
Pin version for first sandbox trial.

iflow-mcp-sixhq-overture

overture-mcp

awaithumans

6Needs review

Candidate set MCP-ready list Skill packages Search

github GitHub

Use this source for discovery, then require profile-level evidence before recommending install or hire.

2listings

Quality82 / 100

Evidence100%

Install ready2

Needs review2

AntonOsika/gpt-engineerHigh traction · MCP ready langgenius/difyHigh traction · MCP ready

Source reviewQuality signals, common risks, install caveats, and trust mix for GitHub.

Strong source signal 82 / 100 source quality

Repository history, stars, forks, issues, and releases can help source review, but they do not prove install safety.

Source metrics

Source quality: 82 / 100
Risk rate: 100%
Evidence coverage: 100%
Average fit: 79 / 100
Install ready: 2
MCP ready: 2
Skills: 0
Needs review: 2
Risk flagged: 2

Common risks

stale maintainers
unreviewed install scripts
unclear release provenance
repository popularity mistaken for trust

Install caveats

Check latest release or commit before copying commands.
Inspect package manager scripts and permission requests.
Prefer tagged releases over arbitrary branches.

AntonOsika/gpt-engineer

langgenius/dify

2High traction

Candidate set MCP-ready list Skill packages Search

awesomeskill Awesomeskill

Use this source for discovery, then require profile-level evidence before recommending install or hire.

1listings

Quality80 / 100

Evidence100%

Install ready1

Needs review1

tiptapHigh traction · MCP ready

Source reviewQuality signals, common risks, install caveats, and trust mix for Awesomeskill.

Strong source signal 80 / 100 source quality

Skill marketplace metadata is useful for discovery, but agents should still verify the underlying source package or manifest.

Source metrics

Source quality: 80 / 100
Risk rate: 100%
Evidence coverage: 100%
Average fit: 70 / 100
Install ready: 1
MCP ready: 1
Skills: 1
Needs review: 1
Risk flagged: 1

Common risks

metadata-only listing
unclear maintainer proof
missing sandbox trial
ambiguous runtime permissions

Install caveats

Fetch profile.json and sourceUrl before use.
Prefer skills with install.json or summary.md.
Run a narrow trial before adding persistent workspace access.

tiptap

1High traction

Candidate set MCP-ready list Skill packages Search

explainx Explainx

Treat this source as exploration only until profile trust, install, evidence, and trial resources are fetched.

1listings

Quality53 / 100

Evidence0%

Install ready1

Needs review1

CrewAINeeds review · MCP ready

Source reviewQuality signals, common risks, install caveats, and trust mix for Explainx.

Needs source review 53 / 100 source quality

This source has weaker standardized provenance. Treat it as a discovery route until profile-level evidence is fetched.

Source metrics

Source quality: 53 / 100
Risk rate: 100%
Evidence coverage: 0%
Average fit: 61 / 100
Install ready: 1
MCP ready: 1
Skills: 0
Needs review: 1
Risk flagged: 1

Common risks

unknown maintainer status
incomplete install metadata
missing community review
unclear update cadence

Install caveats

Fetch trust.json and evidence.json before install.
Use manual review for the first candidate.
Do not infer safety from source label alone.

CrewAI

1Needs review

Candidate set MCP-ready list Skill packages Search

mcpservers Mcpservers

Treat this source as exploration only until profile trust, install, evidence, and trial resources are fetched.

1listings

Quality53 / 100

Evidence0%

Install ready1

Needs review1

FleetQ MCP ServerNeeds review · MCP ready

Source reviewQuality signals, common risks, install caveats, and trust mix for Mcpservers.

Needs source review 53 / 100 source quality

This source has weaker standardized provenance. Treat it as a discovery route until profile-level evidence is fetched.

Source metrics

Source quality: 53 / 100
Risk rate: 100%
Evidence coverage: 0%
Average fit: 61 / 100
Install ready: 1
MCP ready: 1
Skills: 0
Needs review: 1
Risk flagged: 1

Common risks

unknown maintainer status
incomplete install metadata
missing community review
unclear update cadence

Install caveats

Fetch trust.json and evidence.json before install.
Use manual review for the first candidate.
Do not infer safety from source label alone.

FleetQ MCP Server

1Needs review

Candidate set MCP-ready list Skill packages Search

Review support

Review links

Source guidance stays readable; ecosystem evidence, review records, and profile resource order stay linked here.

Open Source guide

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

Source guide Resource map Review guide Site context Agent inventory Sitemap

Source review path

01Site contextShared review rules
02Source guideOpen the focused review record
03Agent inventoryBrowse the matching inventory
04Profile resourcesEvidence, setup, trust, trial, and history files

Use the source data link for complete source-scoped evidence and candidate routes. The HTML page is a compact ecosystem overview; profile resources remain available for finalist review.

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.