Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx agentmods add commands/bengous/claude-code-plugins/skill-reviewgit clone --depth 1 https://github.com/bengous/claude-code-pluginsWhat it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.00012 | $0.01447 |
| Opus 5 | $0.00006 | $0.00724 |
| Sonnet 5 | $0.00002 | $0.00289 |
| Haiku 4.5 | $0.00001 | $0.00145 |
Grade A, and why
skill-review scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 187 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Skill Review
Review a skill against the official skill-creator guidelines using multi-agent consensus evaluation.
Input
$ARGUMENTS - Path to the skill directory to review
Step 1: Load Guidelines
Read the authoritative skill-creator guidelines:
${CLAUDE_PLUGIN_ROOT}/skills/skill-creator/SKILL.md
Extract and remember the key criteria for evaluation.
Step 2: Structural Validation
Run the structural validation script:
"${CLAUDE_PLUGIN_ROOT}/scripts/quick_validate.py" "$ARGUMENTS"
Capture pass/fail status and any error messages.
Step 3: Read Target Skill
- Read
$ARGUMENTS/SKILL.md - Run
ls -la "$ARGUMENTS"to see all files - Run
wc -l "$ARGUMENTS/SKILL.md"to get line count - If references/, scripts/, or assets/ exist, list their contents
Step 4: Multi-Agent Evaluation
Launch 3 parallel Task agents with subagent_type: "Explore". Each evaluates from a different perspective.
CRITICAL: Paste actual content into each agent prompt. Do NOT give file paths and expect agents to read them - this causes agents to evaluate wrong files. Include:
- The relevant skill-creator guidelines sections (pasted, not referenced)
- The target skill's SKILL.md content (pasted in full or excerpted as needed)
- The directory listing (pasted)
Agent 1: Structure & Organization
Prompt must include ACTUAL CONTENT (not file paths):
You are evaluating a skill against the official skill-creator guidelines.
FOCUS ONLY ON:
1. File organization - Does it have SKILL.md? Are scripts/, references/, assets/ properly used?
2. No extraneous files - Are there README.md, CHANGELOG.md, or other junk files?
3. Reference depth - Are references one level deep from SKILL.md (not nested)?
4. Directory naming conventions - Proper structure?
GUIDELINES (excerpt):
[PASTE the "Anatomy of a Skill" and "What to Not Include" sections from skill-creator SKILL.md]
TARGET SKILL NAME: <name>
TARGET SKILL DIRECTORY LISTING:
[PASTE ls -la output here]
TARGET SKILL.md CONTENT:
[PASTE full SKILL.md content here]
For each criterion, rate: PASS / WARN / FAIL
Include brief justification for each rating.
Format as a simple numbered list.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 2d ago First seen · 187 lines · 12 tokens per session scan A 3f18dd9e59f1
skill-review is a command published in the GitHub repository bengous/claude-code-plugins (4 stars, last pushed 2d ago), licensed MIT. It adds 12 tokens to every session and 1,447 once invoked, about $0.0001 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other commands, from other repositories
ai_update_pr
Use the specialized PR Update Expert agent to handle the complete AI-assisted PR workflow with optimized context management.
learn
Extract insights from plan-associated sessions.
spec-init
初始化 OpenSpec (OPSX) 环境 + 验证多模型 MCP 工具.
star-auto
The user has typed star-auto — this file is what that invocation runs. Typing it is the auto grant of conventions §10.7: while this run pursues its goal, the skill each next action names may be started without further naming, the seven explicit-only ones included, and a prepared STOP-line command may be launched…
mcp-builder
Guide for creating high-quality MCP (Model Context Protocol) servers that enable LLMs to interact with external services through well-designed tools. Use when building MCP servers to integrate external APIs or services, whether in Python (FastMCP) or Node/TypeScript (MCP SDK).
xdo
Direct development by the Main Agent.