The Agent Skills Directory

[COMMAND_EXECUTION]: The skill makes extensive use of powerful command-line interfaces to perform its tasks. It uses git for branch management, tagging, and workspace audits; the GitHub CLI (gh) to create releases, upload assets, and monitor CI runs; and macOS security, codesign, and spctl utilities for application signing and notarization. It also invokes package managers such as npm, pnpm, yarn, and bun to build projects and manage dependencies.\n- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted content from the project repository which could be manipulated to influence agent behavior. \n
Ingestion points: The skill reads CHANGELOG.md, package.json, and Git metadata (tags and branch names) as seen in references/general-release-playbooks.md and references/tauri-release-workflow.md.\n
Boundary markers: No explicit delimiters or instructions to ignore embedded commands are used when extracting release notes from the changelog.\n
Capability inventory: The skill possesses full capability to execute shell commands and modify project files.\n
Sanitization: The implementation relies on simple text extraction (e.g., awk) which does not sanitize malicious instructions that might be embedded in the project documentation.\n- [EXTERNAL_DOWNLOADS]: The skill configures and utilizes external GitHub Actions from well-known repositories and interacts with official package registries to execute build pipelines.

lov-release-via-cicd