AI Pieces Collections

A curated collection of reusable AI skills, instructions, and tool playbooks for engineering workflows.

This repository focuses on practical, execution-oriented assets you can plug into Copilot/agent workflows for architecture planning, implementation breakdown, testing, browser automation, documentation, frontend design, and diagram generation.

What is included

• 19 reusable skills under skills/
• 11 reusable instructions under instructions/
• 1 tool playbook under tools/
• 1 MCP integration reference under mcp/
• 1 workspace MCP server config under .vscode/mcp.json
• Supporting reference docs and helper scripts for specific skills

Repository structure

.
|-- .vscode/
|   `-- mcp.json
|-- instructions/
|   |-- agent-skills.instructions.md
|   |-- agents.instructions.md
|   |-- code-review-generic.instructions.md
|   |-- commit-message.instructions.md
|   |-- go.instructions.md
|   |-- kubernetes-deployment-best-practices.instructions.md
|   |-- kubernetes-manifests.instructions.md
|   |-- markdown.instructions.md
|   |-- playwright-python.instructions.md
|   |-- shell.instructions.md
|   `-- sql-sp-generation.instructions.md
|-- mcp/
|   `-- drawio-mcp/
|       `-- README.md
|-- skills/
|   |-- agent-browser/
|   |-- ai-fomo/
|   |-- breakdown-epic-arch/
|   |-- breakdown-feature-implementation/
|   |-- breakdown-test/
|   |-- chrome-devtools/
|   |-- create-architectural-decision-record/
|   |-- create-readme/
|   |-- create-technical-spike/
|   |-- curl-api-testing/
|   |-- documentation-writer/
|   |-- drawio/
|   |-- frontend-design/
|   |-- generate-import-export-import-files/
|   |-- meeting-minutes/
|   |-- playwright-generate-test/
|   |-- playwright-recording/
|   |-- skill-creator/
|   `-- robot-framework-py/
|-- tools/
|   `-- markitdown/
|       `-- README.md
`-- LICENSE

Skill catalog

Skill	Purpose
agent-browser	Install and use the Agent Browser skill for browser-driven automation flows.
ai-fomo	Turn AI information overload into reusable knowledge, signals, and digest workflows.
breakdown-epic-arch	Create high-level technical architecture from an Epic PRD.
breakdown-feature-implementation	Generate detailed feature implementation plans (monorepo-oriented).
breakdown-test	Build test strategy and QA planning artifacts.
chrome-devtools	Browser automation and debugging via Chrome DevTools MCP.
create-architectural-decision-record	Generate structured ADR documents.
create-readme	Create concise, high-quality project README files.
create-technical-spike	Create time-boxed technical spike documents before implementation.
curl-api-testing	Test and debug REST APIs with reproducible curl workflows and compact result reports.
documentation-writer	Produce docs using the Diataxis framework.
drawio	Generate native .drawio diagrams and optional exports.
frontend-design	Build distinctive, production-grade frontend UI and pages.
generate-import-export-import-files	Generate valid/invalid import JSON datasets for import-export testing scenarios.
meeting-minutes	Generate concise meeting minutes with actions and decisions.
playwright-generate-test	Generate Playwright tests from scenario-driven flows.
playwright-recording	Record and refine Playwright tests with codegen workflow.
robot-framework-py	Build/refactor Python-centric Robot Framework test suites.
skill-creator	Create, optimize, and evaluate skills for trigger quality and performance.

Instruction catalog

Instruction	Purpose
agent-skills.instructions.md	Best practices for designing high-quality SKILL files.
agents.instructions.md	Guidelines for creating custom agent files and tool/model settings.
code-review-generic.instructions.md	Reusable code review framework and severity model.
commit-message.instructions.md	Conventional commit format and message quality guidance.
go.instructions.md	Idiomatic Go coding rules and project conventions.
kubernetes-deployment-best-practices.instructions.md	Kubernetes deployment quality and operational best practices.
kubernetes-manifests.instructions.md	Conventions for authoring Kubernetes manifests.
markdown.instructions.md	CommonMark-aligned markdown writing and validation rules.
playwright-python.instructions.md	Playwright + Python test authoring and execution guidance.
shell.instructions.md	Safe, maintainable shell scripting conventions.
sql-sp-generation.instructions.md	SQL stored procedure generation and quality guidance.

Tools catalog

Tool	Purpose
tools/markitdown	Document conversion playbook focused on markdown conversion workflows.

Quick start

1. Clone this repository.
2. Pick an asset type:
   • Skill: skills/<skill-name>/SKILL.md
   • Instruction: instructions/*.instructions.md
   • Tool playbook: tools/<tool-name>/README.md
3. Use the content as your agent instruction/prompt basis.
4. Follow each file's required folder conventions, constraints, and output paths.

Tip

Most skills are designed as opinionated playbooks. Read the full SKILL.md before execution to avoid missing required steps, output format constraints, or expected file locations.

How to use skills effectively

• Keep skills task-specific: one skill per clear outcome.
• Keep instructions language/task specific and scoped by file patterns when possible.
• Preserve each skill's required output format and target path.
• Reuse references/scripts shipped with a skill (for example in skills/robot-framework-py/references/ and skills/robot-framework-py/scripts/).
• For diagram tasks, use the drawio skill and pair it with the MCP reference in mcp/drawio-mcp/ when needed.

MCP server configuration

This repository includes a workspace-level MCP config at .vscode/mcp.json.

Current configured server:

• drawio-mcp-app-server (http)
• URL: http://192.168.137.100:3001/mcp

Example configuration:

{
   "servers": {
      "drawio-mcp-app-server": {
         "url": "http://192.168.137.100:3001/mcp",
         "type": "http"
      }
   },
   "inputs": []
}

Important

If you run this repository in a different environment, update the server URL in .vscode/mcp.json to your reachable MCP endpoint.

Quick check:

1. Ensure the MCP endpoint is running and reachable.
2. Open VS Code in this workspace so .vscode/mcp.json is applied.
3. Use drawio skill flows that depend on MCP connectivity.

Add a new skill

1. Create a new folder under skills/<new-skill-name>/.
2. Add a SKILL.md with:
   • YAML front matter (name, description)
   • clear usage scope
   • strict output requirements
   • examples and constraints
3. Add references/ or scripts/ only when they are truly reusable.
4. Keep instructions deterministic and concise.

Suggested workflow

1. Start from planning skills:
   • breakdown-epic-arch
   • breakdown-feature-implementation
2. Move to quality and delivery:
   • breakdown-test
   • playwright-generate-test or robot-framework-py
3. Document decisions and outcomes:
   • create-architectural-decision-record
   • documentation-writer
   • meeting-minutes

Notes

• License is MIT (see LICENSE).
• MCP reference currently included: drawio-mcp.