source-command-upgrade-deps

Pass

Audited by Gen Agent Trust Hub on Jun 21, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill executes various CLI tools including pnpm (outdated, add, build, test, lint), npm (view), and gh (api, repo, release). These are used to manage the monorepo dependencies and verify the environment after upgrades.
  • [EXTERNAL_DOWNLOADS]: Downloads and installs third-party packages from the NPM registry via pnpm add. It also fetches remote content (changelog files and release notes) from GitHub repositories and general web search results.
  • [PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface by ingesting and analyzing untrusted content from external changelogs to determine migration steps and codebase fixes.
  • Ingestion points: Reads external files and release notes via gh api, mcp__github_repos__get_file_contents, and WebSearch as specified in SKILL.md.
  • Boundary markers: Absent. There are no instructions or delimiters used to help the agent distinguish between the skill's own instructions and potential malicious instructions embedded in retrieved changelogs.
  • Capability inventory: The skill is authorized to install packages (pnpm add) and modify the codebase to fix issues identified during the verification phase (SKILL.md, Step 10).
  • Sanitization: Absent. The skill does not implement any validation, filtering, or escaping of the content fetched from remote sources before processing it.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 21, 2026, 03:25 AM
Security Audit — agent-trust-hub — source-command-upgrade-deps