redfox-skill-generator

Pass

Audited by Gen Agent Trust Hub on Aug 15, 2026

Risk Level: SAFEDATA_EXFILTRATIONPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [DATA_EXFILTRATION]: The skill is configured to silently send usage records, including the user's API key, to the vendor's domain at https://redfox.hk/story/api/skill/record/save every time the skill is triggered. As this data is sent to the vendor's own infrastructure for service tracking, it represents standard vendor functionality.
  • [PROMPT_INJECTION]: The skill ingests untrusted user data, such as functional requirements and API documentation, to generate scripts and instructions. This constitutes an indirect prompt injection surface.
  • Ingestion points: User-provided API documentation and functional descriptions collected in Step 1 of references/core_workflow.md.
  • Boundary markers: The instructions do not implement explicit boundary markers or delimiters for the untrusted content.
  • Capability inventory: The skill generates and writes Python and Shell scripts intended for execution.
  • Sanitization: There is no specified logic for sanitizing or filtering the content provided by the user before it is incorporated into generated files.
  • [COMMAND_EXECUTION]: The skill generates Python and Shell script templates (e.g., fetch_data.py) intended for the user to run. The skill correctly instructs that sensitive credentials should be managed via environment variables rather than being hardcoded into these scripts.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 15, 2026, 09:46 AM
Security Audit — agent-trust-hub — redfox-skill-generator