code-to-spec
Warn
Audited by Gen Agent Trust Hub on Sep 16, 2026
Risk Level: MEDIUMDATA_EXFILTRATIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [DATA_EXFILTRATION]: The skill instructions direct the agent to read and analyze potentially sensitive configuration files. Specifically, Step 2.7 explicitly mentions scanning
.envfiles for environment variables, and Step 9.2 focuses on identifying secrets management patterns. Accessing these files is necessary for the skill's purpose but involves the exposure of sensitive system information and potential credentials contained within the analyzed project. - [INDIRECT_PROMPT_INJECTION]: The skill processes a large volume of untrusted content from the repository being documented, creating a surface for indirect prompt injection where instructions embedded in the code could manipulate the agent's output.
- Ingestion points: Systematic analysis of project structure, entry points, configuration, READMEs, tests, and business logic across all project files (Step 2).
- Boundary markers: The instructions lack specific delimiters or warnings for the agent to ignore embedded instructions found in the analyzed source material.
- Capability inventory: The skill includes the capability to write the resulting specification to the local filesystem (Step 5).
- Sanitization: There are no instructions for sanitizing or escaping external code content before it is synthesized into the final SPEC document.
Audit Metadata