planner
Pass
Audited by Gen Agent Trust Hub on Oct 3, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill functions as a meta-planner that translates project documentation into agent instructions, creating a multi-step execution chain. • Ingestion points: The skill ingests data from local project files including .gsd/SPEC.md and VERIFICATION.md. • Boundary markers: Implementation instructions are strictly encapsulated within and XML-style delimiters. • Capability inventory: The skill defines the ability to generate shell-based verification commands (e.g., npm test, curl) for a downstream agent. • Sanitization: It relies on human-in-the-loop checkpoints (checkpoint:human-verify) and a context-budget-aware planning methodology to prevent instruction drift or quality degradation.
Audit Metadata