hsf

Pass

Audited by Gen Agent Trust Hub on May 16, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill is a developer-focused utility for generating structured specifications. It includes comprehensive documentation and a local validation script.
  • [COMMAND_EXECUTION]: The agent is instructed to use a local Python script (scripts/validate_sesf.py) to validate the syntax of generated documents. This is a functional requirement of the skill and is confined to the skill's own environment.
  • [REMOTE_CODE_EXECUTION]: The scripts/test_validate_sesf.py file contains a subprocess.run call used to test the validator's command-line interface. This is a benign diagnostic function commonly found in software test suites.
  • [PROMPT_INJECTION]: Indirect Prompt Injection Risk. 1. Ingestion points: 'user_request' in SKILL.md. 2. Boundary markers: Absent. 3. Capability inventory: File-read and subprocess calls via 'scripts/validate_sesf.py'. 4. Sanitization: Absent. Malicious instructions in the request could influence spec generation, though impact is limited to the generated document's logic.
Audit Metadata
Risk Level
SAFE
Analyzed
May 16, 2026, 12:01 AM
Security Audit — agent-trust-hub — hsf