setup-skill-discovery
Pass
Audited by Gen Agent Trust Hub on Oct 3, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: Executes
npx skills listandnpx skills searchvia the bash tool to inventory installed skills and find new candidates in the marketplace. - [EXTERNAL_DOWNLOADS]: Recommends the installation of external skill packages from remote repository URLs using the
npx skills addcommand. - [INDIRECT_PROMPT_INJECTION]:
- Ingestion points: Reads content from common repository files such as
README.md,CLAUDE.md,pyproject.toml,package.json,Cargo.toml,Gemfile,Vagrantfile, andDockerfileinSKILL.mdStep 1. - Boundary markers: Absent; the skill instructions do not specify the use of delimiters or directives to ignore instructions within these files.
- Capability inventory: The skill executes bash commands, loads external skill definitions via the
Skill()tool, and writes configuration files to the project directory. - Sanitization: Absent; signals derived from project files are used to construct search queries and recommendations without validating the integrity or safety of the ingested strings.
Audit Metadata