skill-evaluate

A project-aware checker for Claude skills, which are reusable instruction packages for an AI coding assistant. It scans a skill for security issues, compares it with the project, and records a decision.

In plain words
What is it for?
Use it to evaluate a registered skill, local skill folder, GitHub repository, or GitHub URL, then review its benefits, drawbacks, conflicts, and verdict.
Why use it?
It helps you decide whether a skill is safe and relevant before adding it, instead of relying on a feature list or guesswork.

Command for Claude Code

Install

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.

agentmods
npx agentmods add commands/captkernel/skills_curator/skill-evaluate
Clone the repo
git clone --depth 1 https://github.com/captkernel/Skills_Curator

Made for: Claude Code.

Per session 25 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 592 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

What 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.

ModelPer sessionOnce invoked
Fable 5 $0.00025 $0.00592
Opus 5 $0.00013 $0.00296
Sonnet 5 $0.00005 $0.00118
Haiku 4.5 $0.00003 $0.00059

Measured yesterday against content hash 780cc168c4cf, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

skill-evaluate 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 yesterday.

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.

.claude/commands/skill-evaluate.md · 84 lines

How it starts

The opening of the file, as written. The whole thing — 84 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Evaluate a skill

You are evaluating a skill against this project. The user wants a real verdict, not a feature recap. Follow this exact flow.

1. Identify what's being evaluated

$ARGUMENTS is either:

  • A registered skill id (agent-browser)
  • A path to an unregistered local skill folder
  • A owner/repo or full GitHub URL

Resolve which one. If it's a local path, run a security scan first:

python "$HOME/.claude/skills/skills-curator/scripts/registry.py" --check "$ARGUMENTS"

If CRITICAL or HIGH findings appear, stop. Tell the user not to install until each finding is reviewed.

2. Scan the project

python "$HOME/.claude/skills/skills-curator/scripts/registry.py" --scan

Use the project signals (languages, frameworks, goals) as your evaluation lens — does this project genuinely benefit?

3. Read CLAUDE.md and README

Use Read or Glob to find them. Evaluate against what the project says it's building, not against imagined goals.

4. Produce the evaluation in this exact format

## Skill Evaluation: <Name>
Project: <project>
Type: Capability Uplift | Encoded Preference

### ✅ Pros
- <specific, tied to project goals>

### ⚠️ Cons
- <specific cost or limitation>

### 🔴 Conflicts
- <existing skill or pattern that overlaps; "None" if clean>

### 🎯 Verdict: ADOPT | PARTIAL | SKIP
<one or two sentences with the core reason>

### 📦 Adoption Plan
- Adopt: <which features>
- Skip: <which features>
- Pairs with: <skill-id or "nothing">

5. Persist the decision

After the user agrees with the verdict:

python "$HOME/.claude/skills/skills-curator/scripts/registry.py" \
  --eval <id> <project> <verdict> "<summary>" \
  --pros "<a>,<b>" \
  --cons "<c>,<d>" \
  --conflicts "<e>"

Then offer to export the evaluation as a shareable markdown artifact:

python "$HOME/.claude/skills/skills-curator/scripts/registry.py" --export-eval <id>

Why this matters

Other tools install skills. This one persists your judgment so you don't re-decide every time. Treat the registry as the artifact.

Read the full file on GitHub · 84 lines

Changes

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.

  1. yesterday First seen · 84 lines · 25 tokens per session scan A 780cc168c4cf

Subscribe to this mod's changes

skill-evaluate is a command published in the GitHub repository captkernel/Skills_Curator (2 stars, last pushed 9d ago), licensed MIT. It adds 25 tokens to every session and 592 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.