skills/paldom/node-skills/node-ci/Gen Agent Trust Hub

node-ci

Pass

Audited by Gen Agent Trust Hub on Sep 1, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill executes a bundled static analysis script scripts/check_workflows.py to audit generated GitHub Actions workflows for security and hygiene. It also utilizes the GitHub CLI (gh api) to programmatically retrieve commit SHAs, enabling secure version pinning of external actions.
  • [EXTERNAL_DOWNLOADS]: The skill references and downloads metadata from trusted and well-known sources, specifically the official Node.js release schedule on GitHub (nodejs.org) and well-known GitHub Actions provided by the actions/ and pnpm/ organizations. These references adhere to established security guidelines for supply-chain integrity.
  • [INDIRECT_PROMPT_INJECTION]: The skill ingests and analyzes project configuration files, such as package.json and workflow YAML files, to derive CI configurations. While this represents a surface for indirect prompt injection from repository content, the processing is performed via non-executable static analysis and regex-based validation.
  • Ingestion points: .github/workflows/*.yml, package.json (analyzed via check_workflows.py).
  • Boundary markers: None explicitly defined for raw file contents, but the data is processed as structured configuration.
  • Capability inventory: File system writes for workflow creation and execution of the local audit script.
  • Sanitization: The audit script uses specific regular expressions to extract and validate version strings and action identifiers, preventing arbitrary command injection from the analyzed files.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 1, 2026, 06:17 AM
Security Audit — agent-trust-hub — node-ci