github-multi-repo
Warn
Audited by Gen Agent Trust Hub on Feb 20, 2026
Risk Level: MEDIUMEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONREMOTE_CODE_EXECUTION
Full Analysis
- [EXTERNAL_DOWNLOADS] (MEDIUM): The skill utilizes
npx ruv-swarmthroughout multiple examples. This command downloads and executes code from the npm registry. The packageruv-swarmis not from a recognized trusted organization, posing a supply chain risk. - [REMOTE_CODE_EXECUTION] (MEDIUM): In Example 3, the skill clones external repositories and runs
npm test. This is a classic vector for local code execution, as a compromised or malicious repository can define arbitrary commands in its test scripts. - [COMMAND_EXECUTION] (MEDIUM): The skill performs high-impact operations including
git pushandgh pr createacross multiple repositories. While aligned with the stated purpose, this degree of automated write-access to an entire organization requires strictly scoped credentials and carries a high risk of accidental damage or widespread policy violation. - [DATA_EXPOSURE] (LOW): The skill extracts the contents of
package.jsonfrom repositories using the GitHub API and decodes it from Base64. While this data is used for legitimate dependency analysis, it exposes information about the internal structure and dependencies of private repositories to the agent environment.
Audit Metadata