skill-scan
Installation
SKILL.md
Skill Scan (Round 0 pre-screen)
Answers one question fast: "Is this skill safe to install?" It runs a purpose-built scanner over a SKILL.md / skill folder / MCP server and flags the skill-specific threats a normal code read misses: prompt injection hidden in the prose, MCP tool-poisoning, a skill that reaches more than it declares, a skill that reads your credentials, and the classic "reads a secret, posts it to a webhook" payload.
It is a pre-filter. A clean scan does not replace a human review of anything you did not write yourself. A dirty scan stops the install right there.
Recommended review ladder
| Round | Who | What it catches |
|---|---|---|
| 0 | This skill (static scanner) | Known-pattern threats, deterministically, in seconds |
| 1 | A human or agent security review of the source | Intent, architecture, things no pattern matches |
| 2 | A second opinion from a different model family | What round 1's blind spots missed |
Round 0 alone is never sufficient for third-party code.