coding-design-plan
Pass
Audited by Gen Agent Trust Hub on Sep 8, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill instructions require the agent to ingest and analyze untrusted data from the local environment, specifically source code and repository guidance, which serves as an entry point for potential malicious instructions.
- Ingestion points: The 'Workflow' section (Step 1) explicitly directs the agent to 'Inspect the live code, repo guidance, current behavior, and tests'.
- Boundary markers: The skill implements a 'Scope gate' (Step 3) that requires user affirmation or explicit recording of assumptions, and a 'Verification' process (Step 9) to validate the design.
- Capability inventory: No dangerous shell commands, file-system writes, or network operations are directly executed by the instructions in this skill.
- Sanitization: The skill relies on process-based controls such as 'Decision Checks' and 'Scope Discipline' rather than technical input sanitization to manage the analysis of external data.
Audit Metadata