The Agent Skills Directory

[COMMAND_EXECUTION]: The skill discovers and executes various commands found within the repository, including identified 'test gate' commands, release scripts (e.g., scripts/release.*), and package registry publish commands (npm publish, twine upload).
[DYNAMIC_EXECUTION]: The skill implements a workflow that involves identifying arbitrary command strings from CI configurations (GitHub Actions, CircleCI, etc.) and Makefiles, then executing them locally during the release process.
[INDIRECT_PROMPT_INJECTION]: The skill possesses a vulnerability surface for indirect prompt injection through repository metadata and history.
Ingestion points: Reads repository files such as CHANGELOG.md, CI workflow files, and analyzes output from git log to generate release notes.
Boundary markers: The instructions do not specify explicit delimiters or 'ignore' instructions when processing external data from the repository files or logs.
Capability inventory: The skill can write files to the project directory (.omc/RELEASE_RULE.md, .gitignore, CI workflows) and execute discovered shell commands via subprocesses.
Sanitization: No explicit sanitization or validation logic is defined for content extracted from the repository before it is processed or used in decision-making.

release