forge-init

Pass

Audited by Gen Agent Trust Hub on Aug 20, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill performs legitimate project management tasks by generating documentation and maintaining local state files. It does not execute arbitrary shell commands, access sensitive credentials, or perform network requests.
  • [INDIRECT_PROMPT_INJECTION]: The skill ingests user-provided descriptions to generate PRD files, which are then used by other tools. This setup creates a potential surface for indirect prompt injection if the user input contains malicious instructions designed to subvert downstream agent actions. However, as a core functionality of a planning tool, this is considered a standard operational risk. 1. Ingestion points: User-provided description in the /forge-init command. 2. Boundary markers: No explicit sanitization or delimiters defined in the instructions for the generated content. 3. Capability inventory: Writing files to the .rcs/ directory. 4. Sanitization: None specified.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 20, 2026, 11:37 AM
Security Audit — agent-trust-hub — forge-init