breakdown-epic-pm
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests user-provided descriptions of product ideas, which represents an attack surface for indirect prompt injection.
- Ingestion points: The skill uses a
Context Templatethat takes anEpic Ideastring from the user inSKILL.md. - Boundary markers: No specific XML-style delimiters or clear "ignore embedded instructions" warnings are used for the user input.
- Capability inventory: The skill instructs the agent to write a Markdown file to the local file system at
/docs/ways-of-work/plan/{epic-name}/epic.md. - Sanitization: No input validation or escaping is applied to the user input before it is processed by the agent.
- [SAFE]: The skill contains no executable code, remote script downloads, or network requests.
- [SAFE]: The instructions to write documentation to a specific subfolder are aligned with the skill's primary purpose and do not target sensitive system configuration files.
Audit Metadata