The Agent Skills Directory

[COMMAND_EXECUTION]: The skill provides instructions for executing shell commands via Bash, specifically for running project build and test tools like 'pnpm test', 'pnpm build', 'pnpm typecheck', and 'pnpm lint', in addition to standard 'git' operations. While typical for development workflows, this involves the execution of local binaries and scripts based on the skill's guidance.
[INDIRECT_PROMPT_INJECTION]: The skill creates an attack surface where an agent, while attempting to follow these commit guidelines, could be influenced by malicious instructions embedded in the codebase it is analyzing.
Ingestion points: The agent reads local project files and the Git commit history using the 'mcp__git__*' toolset to determine commit content and context.
Boundary markers: The skill does not provide specific delimiters or instructions to ignore potential commands embedded within the files or commit messages it processes.
Capability inventory: The skill allows access to 'Bash' and 'mcp__git__*' tools, enabling the agent to execute shell commands and modify the repository state.
Sanitization: There is no mention of sanitizing or validating the contents of the files or previous commit messages before they are processed by the agent.

clean-commits