api-reference-quality
Pass
Audited by Gen Agent Trust Hub on Oct 26, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to ingest and analyze untrusted external data, including OpenAPI specifications, published documentation pages, gateway logs, and support tickets, to identify documentation gaps. There are no explicit boundary markers (e.g., delimiters) or instructions to ignore embedded commands within these sources, which could lead to the agent being influenced by malicious instructions hidden in the audited data.
- Ingestion points: Sections "1. Census the surface" and "2. Audit every endpoint" in
SKILL.mdspecify reading machine-readable specs, published reference prose, gateway logs, SDK internals, and support tickets. - Boundary markers: None provided to isolate external content from the agent's core instructions.
- Capability inventory: The skill involves complex multi-step analysis and recommends structural changes to CI/CD infrastructure (e.g., linting, contract tests).
- Sanitization: No sanitization or filtering of external content is specified before processing.
Audit Metadata