RP1 `npx` pattern matches across line breaks, so a line ending in "npx" plus the next word becomes a command

[JayOfTheKeyboard]

_RP1_NPX_CMD (mcp_rug_pull.py:124-125) is npx\s+..., and \s matches newlines. So npx at the end of one line and the first word of the next are read as npx <package>. There is also no word boundary before npx.

Minimal repro: a skill whose frontmatter is

---
name: npx
description: repro
---

produces an RP1 finding for npx description (SARIF message: MCP server referenced without pinned version: 'npx\ndescription'). The same happens in wrapped prose, e.g. a line ending "...install it with npx" followed by a line starting "the ...".

Suggested fix: \bnpx[ \t]+(?:-+\w+[ \t]+)*..., keeping the match on one line. Happy to send a PR.

[agentsope]

I opened #646 to address this. The RP1 npx matcher now requires a word boundary and uses spaces/tabs instead of \s, so it cannot combine text across line breaks. Added regressions for the frontmatter and wrapped-prose false positives, identifier suffixes, and normal npx -y commands. The focused MCP rug-pull tests pass (31); hosted test-unit and docker-smoke checks are still running.

[rng1995]

PR-state snapshot checked on 2026-10-04:

• PR #646 — open, non-draft; GitHub review decision: changes requested; latest reported check rollup: success.

Keeping this issue open: the relevant implementation is not merged. Check/review status is a point-in-time snapshot, not a claim of merge readiness.