cowork-plugin-builder

Pass

Audited by Gen Agent Trust Hub on Apr 1, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The documentation includes JavaScript code snippets demonstrating the use of child_process.execSync to execute shell commands. This is presented as instructional content for plugin development.
  • [PROMPT_INJECTION]: The skill outlines a plugin system that ingests external data (e.g., git logs), establishing a surface for indirect prompt injection.
  • Ingestion points: Reading tasks/todo.md, git log output, and pull request data via CLI tools (SKILL.md).
  • Boundary markers: Examples do not demonstrate delimiters or 'ignore' instructions for ingested content.
  • Capability inventory: Example handlers use execSync, fs.writeFileSync, and fs.readFileSync.
  • Sanitization: No input sanitization is shown in the provided templates.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 1, 2026, 05:35 AM
Security Audit — agent-trust-hub — cowork-plugin-builder