source-driven-development
Pass
Audited by Gen Agent Trust Hub on Oct 18, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill's primary function is to fetch and process external documentation, which constitutes an ingestion point for untrusted data that could contain malicious instructions.
- Ingestion points: Official documentation pages fetched from external URLs (e.g., react.dev, docs.djangoproject.com) as described in the 'Fetch Official Documentation' section of SKILL.md.
- Boundary markers: The skill provides conceptual boundaries by instructing the agent to 'Extract only' specific technical information such as API signatures and usage examples, rather than processing the entire page blindly.
- Capability inventory: The skill facilitates code generation and implementation based on fetched content, involving file modification and conversational output.
- Sanitization: The skill includes a 'Retrieval Safety' section that explicitly identifies the risk of prompt injection (LLM01) and directs the agent to ignore instructions targeting the model (e.g., 'ignore previous instructions' or 'reveal system prompt') found within the fetched content.
Audit Metadata