The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes significant local command-line capabilities to diagnose and resolve issues. This includes running git, ssh, ls, and cat commands. It also offers to perform configuration changes like setting credential helpers and generating SSH keys using ssh-keygen and ssh-agent.
[EXTERNAL_DOWNLOADS]: The skill facilitates the installation and use of third-party tools such as BFG Repo-Cleaner, git-secrets via Homebrew, and detect-secrets via PyPI. These are well-known tools for Git maintenance and security.
[PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface in the merge conflict resolution workflow (Workflow 4). Mandatory Evidence: 1. Ingestion points: The agent offers to read the contents of conflicted files to identify markers. 2. Boundary markers: No explicit delimiters or instructions to ignore embedded commands are specified when reading file contents. 3. Capability inventory: The agent has the ability to execute terminal commands (git, rm, etc.). 4. Sanitization: No sanitization logic is present to filter malicious instructions within conflicted files.
[CREDENTIALS_UNSAFE]: The skill interacts with sensitive file paths including the ~/.ssh directory and environment files like .env when helping users locate accidentally committed secrets or setting up authentication.

github-assistant