Production-grade engineering skills for Agent Zero, adapted from addyosmani/agent-skills (MIT).
Nine universal skills covering the full software development lifecycle, plus six reference checklists as supporting material.
| Skill | Trigger |
|---|---|
spec-driven-development |
Start of new project/feature; requirements unclear |
planning-and-task-breakdown |
Before implementation; large or multi-step work |
incremental-implementation |
Building a feature; one slice at a time |
test-driven-development |
Tests as proof; before writing code |
debugging-and-error-recovery |
Tests fail, builds break, behavior unexpected |
code-review-and-quality |
Reviewing a PR, plan, or code change |
code-simplification |
Refactoring for clarity without behavior change |
documentation-and-adrs |
Capturing decisions and rationale |
doubt-driven-development |
Questioning assumptions; avoiding premature commitment |
testing-patterns.mdsecurity-checklist.mdperformance-checklist.mdaccessibility-checklist.mdobservability-checklist.mdorchestration-patterns.md
These are loaded on demand by the relevant skills, not auto-injected.
This plugin is installed at /a0/usr/plugins/agent_engineering_skills/.
Agent Zero auto-discovers skills under usr/plugins/*/skills/ at startup. If the plugin was just added, toggle it off and on in the Plugins dialog (or restart the UI) to trigger a rescan.
skills_tool action=list # via Agent Zero UI or the skills_tool
You should see all 9 skills listed.
python3 /a0/usr/plugins/agent_engineering_skills/scripts/validate-skills.pyThe validator enforces the skill-anatomy contract: every SKILL.md must have valid YAML frontmatter (name and description) and required sections (Overview, When to Use, Process / equivalent, Common Rationalizations, Red Flags, Verification). Warnings are non-fatal; errors must be fixed before shipping a new version.
agent_engineering_skills/
├── plugin.yaml # manifest
├── default_config.yaml # defaults
├── README.md # this file
├── LICENSE # MIT (preserved from upstream)
├── skills/ # 9 SKILL.md files
├── references/ # 6 checklist .md files
└── scripts/
└── validate-skills.py # anatomy validator
These skills are adapted from addyosmani/agent-skills by Addy Osmani (Google Chrome, Lighthouse), MIT licensed.
The original repository ships 26 skills, 4 agent personas, 8 slash commands, 6 references, and validation hooks for Claude Code, Cursor, Gemini CLI, Windsurf, OpenCode, Copilot, Kiro, Codex, and "any agent accepting system prompts or instruction files."
This plugin:
- Adapts 9 of the 26 skills (universal engineering disciplines).
- Drops the platform-specific concepts (slash commands, agents/, hooks/) that don't map to Agent Zero's mechanisms.
- Reuses the validation contract as a Python script (upstream ships Node.js).
- Preserves the substantive content of each skill unchanged — only frontmatter and trigger phrasing were reviewed for Agent Zero's
skills_tooldiscovery model.
The remaining 17 skills (security-and-hardening, performance-optimization, frontend-ui-engineering, api-and-interface-design, git-workflow-and-versioning, ci-cd-and-automation, shipping-and-launch, deprecation-and-migration, observability-and-instrumentation, browser-testing-with-devtools, context-engineering, source-driven-development, idea-refine, interview-me, using-agent-skills) can be added later on request.
MIT. See LICENSE. Copyright (c) 2025 Addy Osmani (original work); adaptation copyright held by this plugin's contributors.
Built and maintained by Shawn5cents (GitHub).
If this plugin saves you time or improves your engineering workflow, you can support its development:
These skills are adapted from addyosmani/agent-skills by Addy Osmani, MIT licensed. See the LICENSE file for full attribution. Original copyright (c) 2025 Addy Osmani; this adaptation copyright (c) 2026 Shawn5cents.