find-reuse
Pass
Audited by Gen Agent Trust Hub on Sep 17, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes local CLI tools (including git, ripgrep, gh, glab, and various package managers) to gather metadata. It explicitly mandates the use of argument arrays (argv) and forbids shell interpolation of queries to prevent command injection vulnerabilities.
- [EXTERNAL_DOWNLOADS]: The skill queries metadata and search results from official and well-known services, including GitHub, GitLab, npmjs, PyPI, crates.io, proxy.golang.org, search.maven.org, nuget.org, rubygems.org, packagist.org, and osv.dev. These operations are limited to metadata retrieval and do not involve executing remote code or scripts.
- [INDIRECT_PROMPT_INJECTION]: The skill has a potential attack surface by ingesting untrusted data from local files (via ripgrep) and external API responses. However, the risk is minimized because the skill is strictly read-only, produces portable reports without mutating the repository, and does not interpolate external data into shell commands. It also enforces a strict input contract (find-reuse.request.v1) to validate requests before execution.
- [PRIVILEGE_ESCALATION]: The skill configuration (references/adapters.json) contains instructions suggesting the use of 'sudo' for tool installation. These are documented as 'install_help' for the user to resolve environment gaps and are not executed by the agent; the skill itself is instructed never to install tools or request administrative credentials.
Audit Metadata