api-patterns

Pass

Audited by Gen Agent Trust Hub on Sep 15, 2026

Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill includes a local script scripts/api_validator.py which is executed to perform static analysis on API-related files and OpenAPI specifications within a provided project path.
  • [INDIRECT_PROMPT_INJECTION]: The skill identifies and processes external data from project files, creating an ingestion surface for indirect instructions.
  • Ingestion points: The scripts/api_validator.py script reads the contents of source files (e.g., .ts, .js, .py) and configuration files (e.g., openapi.json, swagger.yaml) discovered in the target project directory.
  • Boundary markers: No explicit delimiters or instructional constraints are utilized to isolate the untrusted file content from the script's execution context.
  • Capability inventory: The skill is configured with tools for file manipulation (Read, Write, Edit, Glob, Grep) and has the ability to run Python scripts.
  • Sanitization: The script performs no sanitization or filtering of the file contents prior to processing them with regular expressions or JSON parsing.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 15, 2026, 01:29 PM
Security Audit — agent-trust-hub — api-patterns