openpress-apply-comments

Pass

Audited by Gen Agent Trust Hub on Aug 23, 2026

Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
  • [DYNAMIC_EXECUTION]: The skill instructs the agent to use node -e to execute a JavaScript snippet that imports a local module (./packages/core/engine/react/comment-marker.mjs). This is used to programmatically list and decode comment markers from the project's source code.
  • [COMMAND_EXECUTION]: The workflow relies on several shell commands for discovery and verification, including rg (ripgrep) for searching file contents, npm run build for project validation, and node engine/cli.mjs for structural checks and data export.
  • [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and act upon instructions contained within @openpress-comment markers in the project's source files.
  • Ingestion points: The agent reads comment notes and hints from files in the press/ directory via ripgrep and internal engine helpers.
  • Boundary markers: No specific delimiters or "ignore previous instructions" warnings are defined for the content within these markers.
  • Capability inventory: The agent possesses capabilities to modify source files, execute Node.js scripts, and run build commands based on the ingested comment data.
  • Sanitization: The workflow does not explicitly include sanitization or validation of the text within the comment markers before interpreting them as editing instructions.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 23, 2026, 11:10 PM
Security Audit — agent-trust-hub — openpress-apply-comments