cahier-des-charges
Pass
Audited by Gen Agent Trust Hub on Oct 7, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process content from external files located at
.agent/specs/{feature-name}/requirements.mdanddesign.mdto generate its output. - Ingestion points: The workflow in
SKILL.mdexplicitly instructs the agent to read these external markdown files. - Boundary markers: Absent. There are no delimiters or specific instructions provided to the agent to treat content within these files as untrusted data or to ignore any embedded instructions.
- Capability inventory: The skill possesses no dangerous capabilities. It does not have access to network tools, shell execution, or sensitive environment variables. The impact of a successful injection would be limited to the content of the generated document.
- Sanitization: Absent. The skill performs a direct transformation of input data into the output template without validation or filtering.
Audit Metadata