dsh-plugin-lint

Pass

Audited by Gen Agent Trust Hub on Sep 2, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill possesses an attack surface for indirect prompt injection as it is designed to ingest and analyze untrusted third-party plugin artifacts.
  • Ingestion points: The skill reads package.json, cordis.patch.yml, and JavaScript/TypeScript source files from a user-specified directory via the scripts/lint.mjs script and direct file inspection.
  • Boundary markers: There are no explicit boundary markers or 'ignore' instructions defined to prevent the agent from accidentally following instructions embedded in the plugin's metadata or comments.
  • Capability inventory: The skill's environment allows for local file system reads and the execution of the provided Node.js linting script.
  • Sanitization: The linting logic performs structural and banner verification but does not sanitize text content against adversarial prompts that might target the agent during the review process.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 2, 2026, 01:52 PM
Security Audit — agent-trust-hub — dsh-plugin-lint