The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill utilizes npx ruv-swarm to download and execute a remote package from the npm registry for various orchestration tasks such as repository discovery, task execution, and health monitoring.
[COMMAND_EXECUTION]: The skill makes extensive use of system commands including gh (GitHub CLI), git, and npm to perform automated repository operations. These include cloning repositories, creating branches, committing code, and pushing to remote origins.
[DATA_EXFILTRATION]: The skill is designed to read contents from multiple repositories (e.g., via gh api) and programmatically push changes back to GitHub. This represents a data export path that could be misused if the agent is manipulated into handling sensitive information.
[PROMPT_INJECTION]: The skill has a significant attack surface for indirect prompt injection due to its core functionality of processing external repository data.
Ingestion points: The skill ingests untrusted data from external sources using gh repo list (names, descriptions, languages) and gh api (full repository metadata and package.json contents).
Boundary markers: No explicit boundary markers or instructions to ignore instructions embedded in the fetched data are present in the scripts or instructions.
Capability inventory: The agent has access to powerful tools including git push, gh pr create, npm install, and arbitrary command execution via Bash.
Sanitization: The skill lacks sanitization or validation of the fetched metadata or base64-decoded file contents before they are processed by the agent or piped into other tools.

agent-multi-repo-swarm