The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: The skill clones external repositories and executes npm install and npm test within them. This behavior allows for the execution of arbitrary code via package lifecycle hooks or malicious test scripts if a repository being managed is untrusted or has been compromised.
[COMMAND_EXECUTION]: The skill uses a Bash function to execute complex shell commands for repository discovery, cloning, and automation. Examples include using gh repo clone into /tmp/ directories and executing piped logic to automate updates across multiple services.
[EXTERNAL_DOWNLOADS]: The skill uses the GitHub CLI (gh) and npm to fetch source code, API metadata, and package dependencies from remote servers (github.com and npm registries).
[PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection during its coordination and synchronization tasks.
Ingestion points: Data is ingested from external repositories via gh api, gh repo clone, and file reads of package.json and CLAUDE.md.
Boundary markers: There are no explicit delimiters or instructions to ignore embedded commands when the ingested data is passed to the AI swarm agents.
Capability inventory: The skill has the capability to execute shell commands (Bash), push files to GitHub repositories, and manage shared memory.
Sanitization: There is no evidence of sanitization or validation of the content retrieved from external repositories before it is processed by the AI agents.

github-multi-repo