ai-spec-coverage-mapper

Pass

Audited by Gen Agent Trust Hub on Aug 12, 2026

Risk Level: SAFE
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes specification documents and test suite files which are considered untrusted external data. In the provided scripts/ai-coverage.py, this data is interpolated directly into the LLM prompt without explicit boundary markers or sanitization. This creates a surface where malicious instructions hidden within a spec or test file could potentially influence the LLM's mapping output (e.g., a comment in a test file instructing the LLM to report 'full coverage' regardless of actual content).
  • Ingestion points: spec_path and test_globs content (SKILL.md).
  • Boundary markers: None present in the prompt template.
  • Capability inventory: The script executes read operations on the filesystem and network calls to the OpenAI API.
  • Sanitization: No escaping or validation is performed on the ingested file content.
  • [EXTERNAL_DOWNLOADS]: The skill depends on the openai Python package to interact with the LLM service. This is a standard dependency for the skill's stated purpose.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 12, 2026, 10:51 AM
Security Audit — agent-trust-hub — ai-spec-coverage-mapper