agentleFS
caoergoucaoergou/erics-skills/skills/skill-scout/SKILL.md

skill-scout

Skill0 starsChanged 7 months agoMITView on GitHub

Periodic ecosystem radar for AI agent skills. Scans skills.sh, ClawHub, Loaditout, AgentSkillsHub, LobeHub, OpenForge, GitHub and more — produces a shareable digest with security assessments for team evaluation and ecosystem awareness.

24 other projects use a version of this.

Adopted by

Most recently taken up by

Used with

Projects with it also use context-budget.

Other files people use alongside this one.

Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
  • get-proofpilotSkill

    context-budget

    Audits Claude Code context window consumption across agents, skills, MCP servers, and rules. Identifies bloat, redundant components, and produces prioritized token-savings recommendations.

    Also in 3 projects that use this

The file

We split this skill into the rules it gives an agent

5 rules. 0 are common, 4 are distinctive, and 1 is very specific to this project.

  • 0

    Common

    Most of the 23 edited versions we read keep these.

  • 4

    Distinctive

    This author's own choices. Worth a look.

  • 1

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Use broad categories — skills are not limited to coding:

    line 88
  2. Very specific:

    Use mcp__exa__web_search_exa for all web searches (never built-in search)

    line 216
  3. Distinctive:

    Every skill must have a trust tier — no exceptions

    line 218
  4. Distinctive:

    Never recommend tier D; always warn on tier C

    line 219
  5. Distinctive:

    Don't fabricate data — if a source is down, note the gap

    line 220

“Common” is measured: kept by at least 6 in 10 of the 23 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Skill Scout

A periodic ecosystem radar for AI agent skills. Produces a shareable, newsletter-style digest that answers two questions:

  1. What's worth adopting? — Trending skills across all major registries, with security vetting, for team evaluation and selection.
  2. Where is the ecosystem going? — Platform growth, new categories, security landscape, emerging patterns.

The output is designed to be directly shareable — paste into Slack, Lark, a team wiki, or a blog post.

Scope: This skill is for discovery and curation, not for searching by specific requirements. For that, use find-skills.


Workflow

Step 1: Gather — Multi-Registry Scan

Fetch from all major sources in parallel. Each registry surfaces different signals:

SourceWhat to FetchKey Signal
skills.sh/trending pageInstall counts, hot/trending boards (91K+ skills)
ClawHubHomepage + popularStaff picks, download counts (3.2K+ skills)
AgentSkillsHubLeaderboardA-F security grades, heat score (13.8K+ skills)
LoaditoutBrowse pageSecurity grades, MCP + SKILL.md unified (21K+ entries)
LobeHubSkills marketplaceCommunity ratings (100K+ skills)
OpenForgeMarketplaceMulti-framework coverage, DeFi/non-coding categories
GitHubTrending repos with SKILL.mdStars, forks, publisher reputation
WebBlog posts, curated lists, announcementsQualitative signal, community buzz

How:

  • Registry pages: WebFetch each URL
  • Web signal: mcp__exa__web_search_exa — e.g. "AI agent skills trending new noteworthy {current_year}"
  • GitHub: mcp__exa__web_search_exa — e.g. "SKILL.md agent skill site:github.com"

Also monitor well-known publishers for new releases:

PublisherRepoKnown For
anthropicsskillsOfficial reference (82K+ stars)
vercel-labsagent-skillsOfficial Vercel collection (21K+ stars)
antfuskillsCommunity-curated
nicepkgtool-belt, vibe-createUtility & frontend
larksuitelark-skillsLark/Feishu integrations
tech-leads-clubagent-skillsSecurity-validated
gohypergiantagent-skillsEnterprise-grade
openserv-labsskillsMulti-agent workflows

Step 2: Assess — Security Vetting

Every skill in the digest must carry a trust signal. This is non-negotiable.

Why

  • Snyk ToxicSkills (Feb 2026): 36.8% of 3,984 skills had security flaws; 76 confirmed malicious.
  • ClawHavoc: mass malicious upload to ClawHub — credential theft, backdoors, exfiltration.
  • OWASP AST02: Supply Chain Compromise is a top-10 agentic skill risk.
  • Skills have full agent access — filesystem, shell, credentials, network.

Trust Tiers

TierLabelCriteria
SOfficialFrom Anthropic, Vercel, or platform vendor; audited
AHigh TrustKnown publisher, high adoption, security grade A-B, transparent code
BModerateEstablished publisher, decent adoption, no red flags
CCautionNew/low-adoption, ungraded — review SKILL.md before installing
DRiskyUnknown publisher, suspicious patterns — do not recommend

What to Check

  • Publisher: Known org? Account age? Other repos?
  • Registry grade: A-F from AgentSkillsHub or Loaditout?
  • Code: Any curl/wget/eval/exec to unknown URLs? Base64? Encoded payloads?
  • Dependencies: Installs obscure npm/pip packages?
  • Permissions: Requests Bash/network/file access proportional to its purpose?
  • Adoption: Real install counts, stars, community feedback?

Rule: Never recommend tier D. Always warn on tier C. Quick Install section only includes tier A+.

Step 3: Categorize

Use broad categories — skills are not limited to coding:

Development | AI/ML | Productivity | Web & Browser | DevOps & Cloud | Data & Analytics | DeFi & Finance | Marketing & Content | Customer Support | Security | Communication | Science & Research | Education | Design & Media | Blockchain & Web3 | Utility

A skill can span multiple categories. Don't force-fit.

Step 4: Generate Digest

Output the following structure. This is the deliverable — designed to be copy-pasted as-is into Slack, Lark, wiki, or blog.

# Skill Scout Digest

> {date} | Scanned: skills.sh, ClawHub, AgentSkillsHub, Loaditout, LobeHub, OpenForge, GitHub

---

## Ecosystem Pulse

{3-5 sentences on the current state: total skills across registries, growth trend, any platform news, notable shifts in categories or adoption patterns, recent security events}

---

## Recommended Skills

Vetted picks for team adoption. Security tier A or above.

### 1. {skill-name}
> {one-line description}

| | |
|---|---|
| Publisher | {org/repo} |
| Category | {categories} |
| Adoption | {installs} installs / {stars} stars |
| Security | **{tier}** — {one-line justification} |
| Agents | Claude Code, Cursor, Codex, ... |
| Install | `{command}` |

{2-3 sentences: what it does, why it's worth adopting, who benefits}

---

(Repeat for 5-8 top picks across different categories)

---

## Rising — Worth Watching

New or fast-growing skills not yet broadly adopted, but showing momentum:

| Skill | Publisher | Category | Adoption | Security | Why Watch |
|-------|-----------|----------|----------|----------|-----------|
| {name} | {org} | {cat} | {n} | {tier} | {one-line reason} |

---

## Landscape

### By Category

| Category | Trending Skill | Registries Active |
|----------|---------------|-------------------|
| Development | {name} | skills.sh, Loaditout |
| AI / ML | {name} | ClawHub, LobeHub |
| DeFi & Finance | {name} | OpenForge |
| Productivity | {name} | LobeHub, skills.sh |
| ... | ... | ... |

### Platform Health

| Registry | Scale | Security Posture | Trend |
|----------|-------|-----------------|-------|
| skills.sh | {n} skills | No built-in grading | {growing/stable/...} |
| ClawHub | {n} skills | Post-ClawHavoc improvements | {observation} |
| AgentSkillsHub | {n} skills | A-F grading on all entries | {observation} |
| Loaditout | {n} entries | Security grades + MCP support | {observation} |
| LobeHub | {n} skills | Community ratings | {observation} |

---

## Security Radar

### Recent Incidents & Advisories
{Any new supply chain issues, malicious skill takedowns, CVEs, or platform policy changes}

### Install Safety Checklist
1. `uvx mcp-scan@latest --skills` — scan before installing
2. Read the SKILL.md — look for shell commands to unknown URLs, base64, encoded payloads
3. Check the publisher — account age, other repos, community presence
4. Pin versions — use commit SHAs, not `@latest`
5. Least privilege — does the skill need the permissions it requests?

---

## Quick Install

Copy-paste to try the top picks (tier A+ only):

\```bash
{install command 1}
{install command 2}
{install command 3}
\```

---

*Generated by [skill-scout](https://github.com/caoergou/erics-skills) — the ecosystem radar for AI agent skills*

Step 5: Follow-up

---

## Next

1. **Install** — which skill(s) to set up
2. **Deep dive** — fetch full SKILL.md + security review for a specific skill
3. **Compare** — side-by-side on 2-3 skills
4. **Save** — export this digest as a .md file
5. **Audit** — scan your currently installed skills with mcp-scan

Rules

  • Use mcp__exa__web_search_exa for all web searches (never built-in search)
  • Fetch from sources in parallel for speed
  • Every skill must have a trust tier — no exceptions
  • Never recommend tier D; always warn on tier C
  • Don't fabricate data — if a source is down, note the gap
  • Cover all domains, not just coding
  • Note data freshness — when was it last checked
  • For requirement-based search, redirect to find-skills

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.