context-coverage

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

• Third-party content exposure detected (medium risk: 0.30). In scripts/collect.py the org-mode path scan_org_repo() reads repository file listings and context-file line counts via gh api repos/{owner}/{name}/git/trees/... and gh api repos/{owner}/{name}/contents/..., so outsider-authored text in repos’ CLAUDE.md/AGENTS.md/*.rules/skill.md directly becomes LLM-readable content at runtime (line counting), even though the agent does not execute it.