prompt-maker

Pass

Audited by Gen Agent Trust Hub on Aug 27, 2026

Risk Level: SAFE
Full Analysis
  • [PROMPT_INJECTION]: The skill includes instructions in 'rules/context-source-safety.md' to detect and handle adversarial content found in retrieved data. These are defensive guidelines to ensure the agent ignores external instructions that conflict with its primary task.
  • [INDIRECT_PROMPT_INJECTION]: The skill establishes a secure pipeline for processing external repository data. 1. Ingestion points: Repository documents and search tool outputs. 2. Boundary markers: Explicit 'Authority Boundary' rules in the safety guidelines. 3. Capability inventory: A validation script and internal file-read tools. 4. Sanitization: Systematic use of 'Source Ledgers' to track data provenance and trust levels.
  • [COMMAND_EXECUTION]: The skill includes a vendor-specific validation script ('scripts/validate-prompt-maker.mjs'). The script is used locally to verify schema compliance and contains security logic to block access to global or sensitive system directories, such as user home folders.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 27, 2026, 12:29 PM
Security Audit — agent-trust-hub — prompt-maker