ptlam-creating-spec
Pass
Audited by Gen Agent Trust Hub on Sep 2, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from product requirements and feature briefs to produce documentation.
- Ingestion points: SKILL.md (Step 2) directs the agent to read confirmed product scope items and feature briefs which serve as external data sources.
- Boundary markers: The skill provides instructional boundaries by advising the agent to treat PRDs as evidence rather than drafts, but does not define strict character-based delimiters to isolate untrusted content during processing.
- Capability inventory: The agent possesses the ability to create documentation files and directories in a specific path, but is explicitly restricted from overwriting existing files, modifying code, or performing git operations.
- Sanitization: The instructions lack defined procedures for sanitizing or escaping external requirements content before it is interpolated into the generated specification markdown.
Audit Metadata