The Agent Skills Directory

[PROMPT_INJECTION]: The skill possesses an indirect prompt injection surface because it instructs the agent to ingest data from untrusted repository configuration files (such as .commitlintrc or package.json) to determine commit rules. A malicious repository could embed instructions in these files to attempt to influence the agent's logic.
Ingestion points: Project configuration files including .commitlintrc (and its variations) and package.json as described in SKILL.md.
Boundary markers: No delimiters or explicit warnings to ignore embedded instructions are present when reading these files.
Capability inventory: Shell command execution via npx commitlint in SKILL.md.
Sanitization: The skill does not specify any sanitization or validation for the content of the configuration files before they are processed.
[COMMAND_EXECUTION]: The skill executes shell commands using npx commitlint to retrieve configurations and validate commit messages via stdin.
[EXTERNAL_DOWNLOADS]: The use of npx facilitates the execution of the commitlint package, which may involve downloading it from the npm registry if it is not found locally. Commitlint is a well-known community tool for git message linting.

commitlint