Security audit
Security checks for vulnerabilities and agentic risk
This skill coherently analyzes user-selected repositories to create AGENTS.md files, with expected local repository inspection and optional user-supplied reference handling.
Install this skill if you are comfortable letting it inspect the repositories you name and write AGENTS.md-style documentation. Use remote references only from sources you trust, because their AGENTS.md content can influence the generated guidance.
This finding references cloning remote git repositories via subprocess and reading their contents. Fetching and processing attacker-controlled remote repositories can expose the agent to untrusted content, expand data ingress, and potentially trigger unsafe downstream handling if shell arguments, paths, or follow-on tooling are not tightly controlled.
This finding references cloning remote git repositories via subprocess and reading their contents. Fetching and processing attacker-controlled remote repositories can expose the agent to untrusted content, expand data ingress, and potentially trigger unsafe downstream handling if shell arguments, paths, or follow-on tooling are not tightly controlled.
This finding references cloning remote git repositories via subprocess and reading their contents. Fetching and processing attacker-controlled remote repositories can expose the agent to untrusted content, expand data ingress, and potentially trigger unsafe downstream handling if shell arguments, paths, or follow-on tooling are not tightly controlled.
This finding references cloning remote git repositories via subprocess and reading their contents. Fetching and processing attacker-controlled remote repositories can expose the agent to untrusted content, expand data ingress, and potentially trigger unsafe downstream handling if shell arguments, paths, or follow-on tooling are not tightly controlled.
This finding references cloning remote git repositories via subprocess and reading their contents. Fetching and processing attacker-controlled remote repositories can expose the agent to untrusted content, expand data ingress, and potentially trigger unsafe downstream handling if shell arguments, paths, or follow-on tooling are not tightly controlled.
This finding references cloning remote git repositories via subprocess and reading their contents. Fetching and processing attacker-controlled remote repositories can expose the agent to untrusted content, expand data ingress, and potentially trigger unsafe downstream handling if shell arguments, paths, or follow-on tooling are not tightly controlled.
This finding references cloning remote git repositories via subprocess and reading their contents. Fetching and processing attacker-controlled remote repositories can expose the agent to untrusted content, expand data ingress, and potentially trigger unsafe downstream handling if shell arguments, paths, or follow-on tooling are not tightly controlled.
This finding references cloning remote git repositories via subprocess and reading their contents. Fetching and processing attacker-controlled remote repositories can expose the agent to untrusted content, expand data ingress, and potentially trigger unsafe downstream handling if shell arguments, paths, or follow-on tooling are not tightly controlled.
This finding references cloning remote git repositories via subprocess and reading their contents. Fetching and processing attacker-controlled remote repositories can expose the agent to untrusted content, expand data ingress, and potentially trigger unsafe downstream handling if shell arguments, paths, or follow-on tooling are not tightly controlled.
This finding references cloning remote git repositories via subprocess and reading their contents. Fetching and processing attacker-controlled remote repositories can expose the agent to untrusted content, expand data ingress, and potentially trigger unsafe downstream handling if shell arguments, paths, or follow-on tooling are not tightly controlled.
This finding references cloning remote git repositories via subprocess and reading their contents. Fetching and processing attacker-controlled remote repositories can expose the agent to untrusted content, expand data ingress, and potentially trigger unsafe downstream handling if shell arguments, paths, or follow-on tooling are not tightly controlled.
This finding references cloning remote git repositories via subprocess and reading their contents. Fetching and processing attacker-controlled remote repositories can expose the agent to untrusted content, expand data ingress, and potentially trigger unsafe downstream handling if shell arguments, paths, or follow-on tooling are not tightly controlled.
This finding references cloning remote git repositories via subprocess and reading their contents. Fetching and processing attacker-controlled remote repositories can expose the agent to untrusted content, expand data ingress, and potentially trigger unsafe downstream handling if shell arguments, paths, or follow-on tooling are not tightly controlled.
This finding references cloning remote git repositories via subprocess and reading their contents. Fetching and processing attacker-controlled remote repositories can expose the agent to untrusted content, expand data ingress, and potentially trigger unsafe downstream handling if shell arguments, paths, or follow-on tooling are not tightly controlled.
This finding references cloning remote git repositories via subprocess and reading their contents. Fetching and processing attacker-controlled remote repositories can expose the agent to untrusted content, expand data ingress, and potentially trigger unsafe downstream handling if shell arguments, paths, or follow-on tooling are not tightly controlled.
This finding references cloning remote git repositories via subprocess and reading their contents. Fetching and processing attacker-controlled remote repositories can expose the agent to untrusted content, expand data ingress, and potentially trigger unsafe downstream handling if shell arguments, paths, or follow-on tooling are not tightly controlled.
This finding references cloning remote git repositories via subprocess and reading their contents. Fetching and processing attacker-controlled remote repositories can expose the agent to untrusted content, expand data ingress, and potentially trigger unsafe downstream handling if shell arguments, paths, or follow-on tooling are not tightly controlled.
This finding references cloning remote git repositories via subprocess and reading their contents. Fetching and processing attacker-controlled remote repositories can expose the agent to untrusted content, expand data ingress, and potentially trigger unsafe downstream handling if shell arguments, paths, or follow-on tooling are not tightly controlled.
This finding references cloning remote git repositories via subprocess and reading their contents. Fetching and processing attacker-controlled remote repositories can expose the agent to untrusted content, expand data ingress, and potentially trigger unsafe downstream handling if shell arguments, paths, or follow-on tooling are not tightly controlled.
This finding references cloning remote git repositories via subprocess and reading their contents. Fetching and processing attacker-controlled remote repositories can expose the agent to untrusted content, expand data ingress, and potentially trigger unsafe downstream handling if shell arguments, paths, or follow-on tooling are not tightly controlled.
This finding references cloning remote git repositories via subprocess and reading their contents. Fetching and processing attacker-controlled remote repositories can expose the agent to untrusted content, expand data ingress, and potentially trigger unsafe downstream handling if shell arguments, paths, or follow-on tooling are not tightly controlled.
This finding references cloning remote git repositories via subprocess and reading their contents. Fetching and processing attacker-controlled remote repositories can expose the agent to untrusted content, expand data ingress, and potentially trigger unsafe downstream handling if shell arguments, paths, or follow-on tooling are not tightly controlled.
This finding references cloning remote git repositories via subprocess and reading their contents. Fetching and processing attacker-controlled remote repositories can expose the agent to untrusted content, expand data ingress, and potentially trigger unsafe downstream handling if shell arguments, paths, or follow-on tooling are not tightly controlled.
This finding references cloning remote git repositories via subprocess and reading their contents. Fetching and processing attacker-controlled remote repositories can expose the agent to untrusted content, expand data ingress, and potentially trigger unsafe downstream handling if shell arguments, paths, or follow-on tooling are not tightly controlled.
This finding references cloning remote git repositories via subprocess and reading their contents. Fetching and processing attacker-controlled remote repositories can expose the agent to untrusted content, expand data ingress, and potentially trigger unsafe downstream handling if shell arguments, paths, or follow-on tooling are not tightly controlled.
