skill-retro

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

• Third-party content exposure detected (high risk: 0.70). This skill's Step 2 explicitly spawns an analysis sub-agent and instructs it to read the session transcript and the SKILL.md file for each invoked skill (including paths under installed plugin cache such as ~/.claude/plugins/cache/), meaning it ingests third‑party plugin skill files and transcript content that can directly influence analysis and follow-up actions.