api-anything

Pass

Audited by Gen Agent Trust Hub on Mar 14, 2026

Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [PROMPT_INJECTION]: The skill ingests untrusted OpenAPI/Swagger specifications that can influence the behavior of the generated code.
  • Ingestion points: Spec content accepted via URL, local file path, or raw input in SKILL.md Phase 1.
  • Boundary markers: No explicit delimiters are specified for the specification content during processing.
  • Capability inventory: File writes during project generation (Phase 6), network operations for authentication and API calls (Phase 4), and subprocess execution during verification (Phase 7).
  • Sanitization: Implements structure validation in Phase 1 and naming normalization in Phase 2 to ensure canonical command structures.
  • [COMMAND_EXECUTION]: The validation strategy (SKILL.md Phase 7 and references/validation-strategy.md Layer 5) involves building and executing the generated CLI as a subprocess to verify contract compliance.
  • [EXTERNAL_DOWNLOADS]: The ingestion workflow in SKILL.md supports fetching API specifications from user-provided remote URLs.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 14, 2026, 04:06 PM