The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes arbitrary shell commands defined as hooks (prepare_artifact, publish_artifact) within the project's .releaserc.yml configuration file. This allows for project-specific build and publishing logic but also permits the execution of any commands contained in that file.
[COMMAND_EXECUTION]: The skill uses various git and gh (GitHub CLI) commands to perform repository operations, such as creating tags, pushing to remotes, and managing GitHub releases.
[PROMPT_INJECTION]: The skill is subject to indirect prompt injection (Category 8) risks:
Ingestion points: Untrusted data is ingested from git log, commit messages, PR information retrieved via gh pr view, and existing changelog files.
Boundary markers: There are no explicit markers or safety instructions used to delimit this ingested data when the agent processes it to generate new changelog entries or release notes.
Capability inventory: The skill has the capability to execute shell commands (git, gh, and hooks), write to the local filesystem (version files, changelogs), and interact with GitHub APIs.
Sanitization: The skill follows the best practice of using temporary files for release notes to prevent shell command injection, but it does not specify content sanitization to prevent the agent from obeying instructions that might be embedded in commit messages or PR titles.

release-skills