forge-offline

Pass

Audited by Gen Agent Trust Hub on Aug 6, 2026

Risk Level: SAFECOMMAND_EXECUTIONDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill executes project-specific CLI tools including forge, fullstack-forge, and detect-stack to perform its auditing functions.
  • [DATA_EXFILTRATION]: The skill references a file at ../fullstack-forge/references/PROTOCOL.md, which involves a relative path traversal outside the standard repository root. This is used to read configuration from the associated 'Fullstack Forge' bundle.
  • [PROMPT_INJECTION]: The skill identifies an indirect prompt injection surface through the ingestion of external data from project files such as PROTOCOL.md and .forge/project-profile.json.
  • Ingestion points: Reads project configuration files and documentation from the local filesystem.
  • Boundary markers: No explicit instructions to ignore embedded commands within the ingested files are present.
  • Capability inventory: The skill can execute local CLI tools and modify files in its fix mode.
  • Sanitization: There is no explicit mention of sanitizing or escaping the content read from the configuration files before processing.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 6, 2026, 06:48 PM
Security Audit — agent-trust-hub — forge-offline