token-coach

Pass

Audited by Gen Agent Trust Hub on May 2, 2026

Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill uses shell commands to resolve the file path for measure.py (a script from the companion token-optimizer skill) and executes it using python3. This execution is used to retrieve JSON-formatted token usage snapshots, quality metrics, and to perform setup tasks like 'Smart Compaction'.
  • [REMOTE_CODE_EXECUTION]: The skill performs dynamic execution of a Python script located at a path computed at runtime. While this targets a local script within the Claude Code environment rather than a remote download, it represents a cross-skill dependency where the logic of token-coach relies on the external code provided by the token-optimizer package.
  • [INDIRECT_PROMPT_INJECTION]: The skill ingests and parses JSON data produced by the measure.py script. This represents a potential injection surface if the metrics script were to output untrusted content, though in this context, it is a standard data-processing flow for the agent to provide informed advice.
Audit Metadata
Risk Level
SAFE
Analyzed
May 2, 2026, 03:30 PM