clarification
Pass
Audited by Gen Agent Trust Hub on Sep 24, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill includes an inherent surface for indirect prompt injection because it interpolates user-supplied request summaries directly into subagent prompts. 1. Ingestion points: User requests are ingested through the frontmatter triggers and processed in the Q&A loop. 2. Boundary markers: The subagent prompt template in the 'Track 2' section does not employ explicit delimiters (e.g., XML tags or triple quotes) around the interpolated user request. 3. Capability inventory: The skill is designed to dispatch 'Explore' subagents for reading the codebase and can write Markdown context briefs to the local file system. 4. Sanitization: No explicit sanitization or filtering logic is provided for the user-supplied content before it is passed to subagents.
- [SAFE]: The skill's operations are confined to the intended purpose of requirement clarification. It includes internal safety gates, such as recommending against starting implementation before ambiguity is resolved and requiring user confirmation of the final Context Brief. No signs of obfuscation, credential exfiltration, or unauthorized command execution were found.
Audit Metadata