skills
gitlabProduction-ready agent skills, automation recipes, and developer workflows for AI coding assistants.
https://gitlab.com/bitzzdev/skills.git- Python98.7%
- Shell1.3%
skills
A collection of agent skills for Antigravity, Gemini CLI, Claude Code, OpenCode, and Claude.ai. Each skill is a self-contained folder with a SKILL.md and reference files that load only when the agent needs them.
Skills
| Skill | What it does |
|---|---|
android-app-dev | Builds, edits, and debugs native Android apps in Kotlin or Java, with Jetpack Compose or XML views, Gradle setup, and GitHub Actions CI builds. |
context-rich-commits | Turns git history into project memory. Commit messages record why, what changed, decisions, and gotchas; checkpoint commits snapshot the repository state. Includes a commit linter and history reader. |
credential-safety | Checks worktree files, commit ranges, ignore rules, and scanner results so credentials do not reach remote repositories. |
deep-research | Investigates source and target repositories before copying or porting a feature, tracing behavior, integration points, licensing, and validation. |
debug-and-fix | Finds and fixes bugs end-to-end. Reproduces the error, searches affected files, checks full logs, reads official documentation, fixes root causes, and verifies. Includes error context extraction scripts. |
error-docs-lookup | Directs the agent to read current official documentation, changelogs, and GitHub issues before fixing build, type, lint, install, or runtime errors instead of guessing from memory. |
google-seo-skill | Applies official Google SEO and generative AI search (AEO/GEO) guidance when writing, reviewing, or auditing web pages. Used for public web projects needing search ranking; avoided for personal or internal projects where indexing is not needed. |
human-writing | Keeps written output sounding human. Bans em dashes and emoji, eliminates stock AI phrasing, and covers code comments, commit messages, READMEs, UI copy, and technical social posts on X and Reddit. Includes a scanner script. |
iris-shaderpack-dev | Creates and debugs Minecraft shaderpacks for Iris and OptiFine, including shaders.properties, program files, and common shader effects. |
minecraft-fabric-mod | Scaffolds and updates Fabric mods for Minecraft Java Edition, with version-sensitive Gradle, Loom, mapping guidance, and CI workflows. |
persistent-goal-execution | Keeps agents focused on the complete requested outcome across long tasks, retries, waiting, context compaction, and failures. |
production-ready | Takes a project to a production release. Directs the agent to orchestrate all other relevant skills in its toolkit, audits the repository, generates changelogs and docs, drafts legal files, and conducts a multi-skill refresh pass. Includes an audit script. |
threejs-awwwards-site | Builds Three.js and WebGL websites with scroll-driven motion, smooth scrolling, and cinematic transitions. |
Skill Orchestration
The skills work together as an interconnected toolkit:
- When preparing a release, invoke
production-ready. It acts as the orchestrator, detecting the project type and runningpersistent-goal-execution,debug-and-fix,error-docs-lookup,google-seo-skill(if a public website needing search ranking),human-writing, andcontext-rich-commits. - Before committing or pushing a release,
production-readyrunscredential-safetyto check for exposed credentials and stop unsafe pushes. - When a release includes a feature copied or ported from another repository,
production-readyrunsdeep-researchbefore accepting the implementation. production-readypushes every release todevfirst and promotes the exact commit to production only after itsdevCI/CD check passes.- For personal or internal projects where search indexing is not desired,
production-readyautomatically avoids SEO ranking checks and ensuresnoindexor authentication guards are in place. production-readyverifies the corporate release document suite (terms, privacy, licenses, third-party attribution, security policies, and community standards) while properly handling private projects, user exclusions, and release overrides.- When an error occurs during any build or test,
debug-and-fixautomatically pairs witherror-docs-lookupto consult official live documentation before making changes. - For difficult or multi-step work,
persistent-goal-executionkeeps the full objective, acceptance criteria, current plan, authoritative evidence, and completion state active until the requested end state is proven. - Every skill includes an update check directing agents to inspect git status in the skill directory (such as
~/.copilot/skills/or~/.gemini/antigravity-cli/skills/) before running, so instructions stay current.
Installation
GitHub Copilot
Copy the skill folder into your personal Copilot skills directory:
git clone https://github.com/bitzzdev/skills.git
cp -r skills/google-seo-skill ~/.copilot/skills/
Or clone the entire repository directly into ~/.copilot/skills/.
Antigravity and Gemini CLI
Copy the skill folder into your personal skills directory:
git clone https://github.com/bitzzdev/skills.git
cp -r skills/google-seo-skill ~/.gemini/antigravity-cli/skills/
For project-specific skills, place them in .gemini/skills/ within your project root.
Claude Code
Copy the skill folder into your personal skills directory, or into a project:
cp -r skills/google-seo-skill ~/.claude/skills/
For a single project, use .claude/skills/ inside that project instead.
OpenCode
Copy the skill folder into your OpenCode skills directory (such as ~/.config/opencode/skills/).
Claude.ai
Zip the skill folder so SKILL.md sits inside a top-level folder named after the skill, rename the archive to <skill-name>.skill, and upload it from the skills section of your Claude.ai settings.
Layout
skill-name/
SKILL.md required, YAML frontmatter plus instructions
references/ optional, loaded when the skill points to them
scripts/ optional, helpers the skill can run
The name in the frontmatter must match the folder name. A GitHub Action checks this on every push and pull request. You can run the same check locally:
pip install pyyaml
python scripts/validate_skills.py
Contributing
Fixes and new skills are welcome. See CONTRIBUTING.md.
License
MIT. See LICENSE.
Support
If these are useful to you, you can support the work on Ko-fi.