executing-roadmap-checkpoints
Pass
Audited by Gen Agent Trust Hub on Sep 1, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted content from the local repository which could contain malicious instructions designed to influence agent behavior.
- Ingestion points: Reads roadmap markdown files (
docs/<slug>/*-roadmap.md), state files (docs/<slug>/roadmap-state.json), and project configuration files (CLAUDE.md,AGENTS.md,.claude/rules). - Boundary markers: The instructions do not define clear delimiters or "ignore instructions" directives when parsing the content of these external files.
- Capability inventory: The skill has the authority to read, write, and delete files (cleanup of plans and roadmaps) and can trigger powerful sub-agent skills like
subagent-driven-developmentandcodex-implementwhich perform code modification. - Sanitization: There is no explicit sanitization or validation of the roadmap headings or dependency IDs, which are used to derive file paths for deletion (e.g., in
docs/<slug>/steps/<step-slug>/).
Audit Metadata