bep-best-openapi-design
Pass
Audited by Gen Agent Trust Hub on Sep 3, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to execute a bundled Python script (
scripts/check_resource_paths.py) to validate API paths. - Evidence: The script uses only Python standard libraries (
argparse,re,sys,pathlib,urllib.parse) and performs local path analysis without network or sensitive file access. - [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data like OpenAPI descriptions and source code to generate design recommendations or implementation code.
- Ingestion points: Reads repository instructions, existing routes, OpenAPI descriptions, schemas, clients, tests, and API conventions to inform its design decisions.
- Boundary markers: Not explicitly defined for external content, though the skill enforces a strict "fail closed" policy for incomplete resource models.
- Capability inventory: The skill generates OpenAPI contracts, implementation code, and executes a local path-checking script.
- Sanitization: No specific sanitization or filtering of external content is detailed before processing.
Audit Metadata