release-publish

Pass

Audited by Gen Agent Trust Hub on Mar 12, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill mandates the execution of compose-agentsmd at the start of every session to refresh agent rules and provides instructions for using sudo directly when elevated privileges are necessary for system-level operations.
  • [EXTERNAL_DOWNLOADS]: The agent is directed to install several NPM packages globally, including compose-agentsmd, @metyatech/task-tracker, and @metyatech/thread-inbox, to manage tasks and rules within the local environment.
  • [REMOTE_CODE_EXECUTION]: The skill dynamically populates its operational rules (AGENTS.md) by fetching content from the metyatech/agent-rules repository on GitHub.
  • [PROMPT_INJECTION]: The skill possesses an indirect prompt injection surface; it ingests untrusted data from repository files like CHANGELOG.md and metadata (ingestion points) without explicit boundary markers or sanitization, while maintaining high-impact capabilities (capability inventory) such as npm publish, git push, and sudo execution.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 12, 2026, 02:10 AM