skill-manager

Pass

Audited by Gen Agent Trust Hub on May 1, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill instructs the agent to run a local Python script (analyze_project.py) located in the skill's internal directory to automate the process of stack detection and scoring of other installed skills.
  • [PROMPT_INJECTION]: The skill's primary function is to inject instructions into the CLAUDE.md file that direct the agent to "ignore" or "not invoke" specific skills. This is a functional use of instructions to modify and limit the agent's available capabilities for the current project context.
  • [DATA_EXFILTRATION]: The agent reads the global configuration file ~/.agents/.skill-lock.json to obtain a list of all installed skills. This access is necessary for the skill to perform its intended management functions across all project-specific contexts.
Audit Metadata
Risk Level
SAFE
Analyzed
May 1, 2026, 03:01 AM
Security Audit — agent-trust-hub — skill-manager