The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process untrusted data from git commit messages, pull request titles, and pull request descriptions to generate automated documentation.
Ingestion points: Commit history and pull request metadata processed by semantic-release, standard-version, and commitizen (as described in resources/implementation-playbook.md).
Boundary markers: Absent. The provided templates interpolate raw commit data directly into Markdown files and GitHub Release descriptions.
Capability inventory: The tools configured by this skill perform file system writes (CHANGELOG.md), Git operations (commit/push), and external API calls (creating GitHub Releases).
Sanitization: Absent. There is no explicit logic provided to sanitize or validate the content of commit messages before inclusion in the generated changelog.
[EXTERNAL_DOWNLOADS]: The skill instructs the user to download and install several third-party software packages and GitHub Actions.
Node.js Packages: Fetches @commitlint/cli, @commitlint/config-conventional, husky, standard-version, and semantic-release from the npm registry.
Python Packages: Fetches commitizen from PyPI.
GitHub Actions: The CI/CD template references actions/checkout@v4, actions/setup-node@v4, and the third-party softprops/action-gh-release@v1 action.
[COMMAND_EXECUTION]: The skill provides numerous shell commands and configuration scripts that execute localized logic.
Patterns: Includes scripts for initializing husky, configuring commitlint, and running versioning tools that execute Git commands and file manipulations.
Execution surface: Deployment workflows in resources/implementation-playbook.md involve executing Node.js binaries and Git CLI operations within a CI environment.

changelog-automation