hooks — independently scanned and version-tracked by SaferSkills.
SaferSkills independently audited hooks (Hook) and scored it 100/100 (green). The audit ran 55 deterministic rules across Security, Supply Chain, Maintenance, Transparency, and Community; it found 0 high-severity and 0 lower-severity findings. The full rule-by-rule trace and per-finding evidence are below. Free, methodology-open.
Findings & checks · 0 flagged
Every scanned point with the score it earned and what moved between them.
First recorded scan — no prior version to compare against.
The primary manifest — the file an agent reads to learn what this artifact does.
A Claude Code skill that forces AI to understand before executing.
AI answer quality depends on context quality, not prompt decoration.
A UserPromptSubmit hook that automatically intercepts complex requests and injects two disciplines:
Simple commands (git status, read file.py, /commit) pass through untouched.
When your brain is foggy, you write bad prompts, get bad answers, and spiral. This hook acts as a guardrail — it forces AI to pause, restate, search the relevant domain, and confirm direction before diving in. The core insight: most bad AI answers come from missing context and skipped prior art, not missing prompt tricks.
In your terminal:
claude plugin marketplace add recomby-ai/promptly-prompt
claude plugin install promptly-prompt@promptly-promptOr inside a Claude Code session, prefix with !:
! claude plugin marketplace add recomby-ai/promptly-prompt
! claude plugin install promptly-prompt@promptly-promptclaude plugin uninstall promptly-prompt@promptly-prompt
claude plugin marketplace remove promptly-promptThe hook script scores every prompt's complexity using rule-based signals:
| Signal | Score |
|---|---|
| Long text (>80 words / chars for Chinese) | +2 |
| Medium text (40–80 words) | +1 |
| Multiple sentences (>3) | +1 |
| Multi-step words — each match ("then", "然后", …), cap +3 | +2 each |
| Ambiguity words — each match ("maybe", "大概", …), cap +3 | +1 each |
Multiple questions (>1 ?) | +1 |
| Code blocks / file paths | −2 |
| Imperative verb at start ("fix", "run", "git") | −1 |
Slash command (/commit) | −3 |
| Very short (<10 chars) | −3 |
Score ≥ 3 triggers injection. Below 3 passes through silently.
No API calls. No dependencies beyond Python stdlib. Runs in < 50ms.
Inspired by promptly-prompt, a prompt optimization project built during sophomore year. The original vision was a full prompt optimization pipeline — this skill distills the core insight: understand first, find existing methods, then act.
From prompt engineering to context engineering.
MIT
~30 seconds. Free. No account. Every finding cites a rule and a line of evidence.