npm-package-publishing

Pass

Audited by Gen Agent Trust Hub on Jul 22, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill references several third-party security auditing and maintenance tools.
  • Evidence in SKILL.md: References tools such as open-packages-on-npm, zizmor, publint, and multiocular from various community sources.
  • These references are intended for legitimate security auditing and enhancement tasks.
  • [REMOTE_CODE_EXECUTION]: The skill recommends executing several auditing tools via npx, which involves downloading and running code from the npm registry.
  • Evidence in SKILL.md: Commands include npx @e18e/setup-publish, npx actions-up, and npx publint.
  • These operations are standard practices for JavaScript project maintenance and auditing.
  • [INDIRECT_PROMPT_INJECTION]: The skill ingests and audits untrusted project configuration files, representing a potential surface for indirect prompt injection.
  • Ingestion points: The agent is instructed to inspect package.json, lockfiles, .npmrc, and GitHub workflow files (SKILL.md, Agent Workflow section).
  • Boundary markers: None are specified for the file ingestion process.
  • Capability inventory: The skill can modify repository files and run local validation tools (SKILL.md, Responsibility Split section).
  • Sanitization: No explicit sanitization or filtering of the ingested file content is mentioned.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 22, 2026, 07:27 PM
Security Audit — agent-trust-hub — npm-package-publishing