gskills-forge

Pass

Audited by Gen Agent Trust Hub on Jun 22, 2026

Risk Level: SAFE
Full Analysis
  • [COMMAND_EXECUTION]: The skill executes 'gemini-forge' via package.json scripts and procedural instructions (e.g., 'gemini-forge check .'). This utility is attributed to the skill's author ('tercel') and is used for deployment and validation within the local environment.
  • [DATA_EXPOSURE]: The skill performs directory listings ('list_directory') and file reads ('read_file') to audit and refactor existing skills within the workspace ('.gemini/skills/' or '.agents/skills/'). This access is standard for skill management tasks.
  • [DYNAMIC_EXECUTION]: The '/new' and '/convert' commands generate new file structures and SKILL.md content based on established templates and user-provided inputs. This template-based generation is a standard feature for a scaffolding and migration utility.
  • [INDIRECT_PROMPT_INJECTION]: The '/audit' and '/convert' commands ingest external skill files and instructions for processing. While this creates a potential attack surface for indirect prompt injection, the risk is mitigated by the skill's specific focus on architectural analysis and transformation rather than unvalidated execution.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 22, 2026, 12:49 PM
Security Audit — agent-trust-hub — gskills-forge