shunk031-doc-slop-review

Pass

Audited by Gen Agent Trust Hub on Aug 25, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [COMMAND_EXECUTION]: The skill uses subprocess.run to execute command-line tools including git, npx, and codex. These executions are well-scoped and essential for the skill's purpose of linting documents and managing isolation environments.
  • [EXTERNAL_DOWNLOADS]: The review script uses npx to fetch and run pinned versions of textlint and its plugins from the official npm registry. These are well-known developer tools and the downloads are handled by a standard package runner.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted user-supplied documents as input for an LLM judge. It mitigates injection risks by using clear delimiters and explicit instructions to the model to ignore any commands or instructions contained within the document text.
  • [DATA_EXPOSURE]: The skill implements isolation by copying local Codex configuration and authentication files into a temporary directory for the duration of the review. This is designed to prevent the LLM judge from accessing sensitive project context or history from the main environment.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 25, 2026, 01:14 PM
Security Audit — agent-trust-hub — shunk031-doc-slop-review