v-implementing
Pass
Audited by Gen Agent Trust Hub on Oct 8, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill reads and interprets the structure of plan files (root.md and step-.md) to schedule and implement tasks.
- Ingestion points: Frontmatter and content of local markdown files generated by a planning process.
- Boundary markers: None; plan files are used as the primary source of implementation instructions.
- Capability inventory: Orchestrates sub-agents using the Agent tool, performs git operations (branching, commits), and executes background verification scripts.
- Sanitization: No external data sanitization is performed on the plan files, as they are considered trusted local configuration for the implementation workflow.
Audit Metadata