Skills
skills/arabelatso/skills-4-se/specification-driven-generation/Gen Agent Trust Hub

specification-driven-generation

Pass

Audited by Gen Agent Trust Hub on Mar 6, 2026

Risk Level: SAFE
Full Analysis
  • [PROMPT_INJECTION]: The skill processes untrusted user-provided specifications, which represents a surface for indirect prompt injection. This behavior is necessary for its function as a code generator. 1. Ingestion points: Natural language descriptions and API specifications (SKILL.md). 2. Boundary markers: Not present. 3. Capability inventory: Generates source code and suggests running unit tests via shell commands. 4. Sanitization: Not present.
  • [COMMAND_EXECUTION]: The skill instructions include the execution of standard testing commands to verify implementation quality. Evidence: Suggested use of pytest and mvn test in the verification section.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 6, 2026, 10:21 PM