audit-env-parity

Pass

Audited by Gen Agent Trust Hub on Aug 25, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill functions as a read-only auditor for environment configuration parity, assisting in identifying missing variables or naming drifts across development, staging, and production environments without modifying the system state.
  • [DATA_EXPOSURE]: The skill analyzes sensitive configuration sources including .env files and environment variables. It maintains safety through explicit directives: 'Never print secret values', 'Diff names only (never values)', and 'Three-way gap produced with no secret values'.
  • [INDIRECT_PROMPT_INJECTION]: The skill ingests data from external project code and configuration files. It addresses injection risks through the following evidence chain:
  • Ingestion points: Reads .env* files, .env.example, and code patterns like process.env.X (SKILL.md).
  • Boundary markers: Explicitly mandates a 'no secret values' policy and the use of parity matrices to isolate data.
  • Capability inventory: Requires file system read and environment variable access to perform comparisons.
  • Sanitization: Limits the output to variable names and presence/absence status, filtering out the actual content of the secrets.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 25, 2026, 05:55 AM
Security Audit — agent-trust-hub — audit-env-parity