ln-622-build-auditor

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

• Third-party content exposure detected (high risk: 1.00). SKILL.md explicitly instructs fetching missing reference files via WebFetch from the public URL https://raw.githubusercontent.com/levnikolaevich/claude-code-skills/master/skills/{path} and mandates reading those shared/references (e.g., ci_tool_detection.md), which the agent parses and uses to decide build commands and severity/scoring—so untrusted public GitHub content can materially influence tool use and decisions.

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

• Potentially malicious external URL detected (high risk: 1.00). The skill explicitly states that if local shared/ is missing it will WebFetch required "MANDATORY READ" files from https://raw.githubusercontent.com/levnikolaevich/claude-code-skills/master/skills/{path}, and those fetched markdowns are required at runtime and directly control the agent's instructions and behavior.