pitch-gen
Pass
Audited by Gen Agent Trust Hub on Aug 26, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted user input in the form of a startup idea and interpolates it directly into an LLM prompt without sanitization or boundary markers.
- Ingestion points: The
ideaargument captured insrc/cli.tsis passed directly to thegeneratePitchfunction insrc/index.ts. - Boundary markers: There are no delimiters or explicit instructions to the model to ignore embedded commands within the user-supplied content in
src/index.ts. - Capability inventory: The skill performs network requests to the OpenAI API and possesses the capability to write the generated response to the local filesystem via
fs.writeFileSyncinsrc/cli.ts. - Sanitization: The input string is not validated or escaped before interpolation into the prompt template.
Audit Metadata