cli-anything

CRITICAL E005: Suspicious download URL detected in skill instructions.

• Suspicious download URL detected (high risk: 0.70). One URL is a direct remote installer shell script (astral.sh/uv/install.sh) and another is a personal GitHub repo used as a direct install target (akillness/jeo-skills) — both are high-risk delivery vectors for arbitrary code and should be treated as suspicious until verified.

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (high risk: 0.70). At runtime, the required workflow for harness generation/refinement can fetch and ingest outsider-authored content from public sources (e.g., GitHub repos or other registries via /cli-anything <path-or-repo> / cli-hub install), whose README/issues/code text may be parsed into the LLM’s context during the 7-phase pipeline.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.80). The skill contains explicit runtime commands that fetch and execute remote repositories — e.g. "git clone https://github.com/HKUDS/CLI-Anything.git && bash CLI-Anything/codex-skill/scripts/install.sh" and npx installs from "https://github.com/akillness/jeo-skills", which will retrieve and run remote content.