ce-compound-refresh

Pass

Audited by Gen Agent Trust Hub on Jun 23, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill's instructions and supporting files demonstrate a legitimate purpose for managing technical debt in documentation. It follows security best practices such as using subagents for context isolation when processing untrusted content.
  • [COMMAND_EXECUTION]: The skill performs standard Git operations (branching, staging, committing, and opening PRs) and executes a local Python script (validate-frontmatter.py) for YAML validation. These commands are gated by user consent in interactive mode and are essential for the skill's functionality.
  • [DATA_EXFILTRATION]: No evidence of unauthorized data exfiltration was found. The skill processes project-specific documentation and interacts with official Git platforms as part of the intended developer workflow.
  • [PROMPT_INJECTION]: The instructions do not contain attempts to override agent safety guidelines or bypass constraints. It uses structured interaction patterns (AskUserQuestion) to manage user engagement safely.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 23, 2026, 09:14 PM
Security Audit — agent-trust-hub — ce-compound-refresh