ecosystem-health

Pass

Audited by Gen Agent Trust Hub on Mar 17, 2026

Risk Level: SAFE
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill retrieves the official Claude Code CHANGELOG and documentation from the Anthropics GitHub repository via the docs-management skill. This access targets a trusted organization and is used for version and health tracking.
  • [COMMAND_EXECUTION]: It orchestrates audits by running local commands and specialized sub-agent skills (e.g., /audit-skills). These actions are restricted to the local environment and are consistent with the skill's management functions.
  • [DATA_EXFILTRATION]: The skill maintains a local configuration file (.claude/ecosystem-health.yaml) for state persistence. It does not access or exfiltrate sensitive user credentials or private information.
  • [PROMPT_INJECTION]: It processes documentation content to automate health checks. The risk of indirect prompt injection is mitigated by sourcing content from trusted repositories and using structured validation logic.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 17, 2026, 10:32 AM
Security Audit — agent-trust-hub — ecosystem-health