Agent Skills Collection

A curated collection of Agent Skills — modular knowledge packs that teach AI coding assistants domain-specific conventions, best practices, and architectural patterns. Each skill is self-contained and can be installed independently into any project.

---

Available Skills

More skills coming soon. This collection is actively growing — check back for new additions or contribute your own.

---

📦 SAP Commerce Cloud Development

This skill teaches AI agents how to work with SAP Commerce Cloud (Hybris) backend codebases.

Highlights

• 🏗️ Architect extensions using the proper layered pattern (Type System → Services → Facades → OCC API)
• 📋 Define types in items.xml with correct deployment, typecodes, and modifiers
• 🔧 Configure Spring beans, aliases, and overrides in *-spring.xml
• 🌐 Build OCC REST endpoints with controllers, WsDTOs, and DataMapper
• 📦 Manage data through the OOTB patches framework and ImpEx
• ✅ Write tests following the AAA pattern with JUnit 5 and Mockito
• 🚀 Build & deploy using the correct Ant lifecycle commands

Key Conventions Enforced

Reference Docs

• Type System — items.xml: types, enums, relations, deployment
• Patches & ImpEx — Patches framework, ImpEx syntax, macros, lifecycle
• Spring Configuration — Beans, aliases, overrides, list/map merge
• OCC REST API — Controllers, WsDTOs, beans.xml, DataMapper
• Testing — Unit tests, Mockito, AAA pattern

Agent Compatibility

This skill is compatible with any agent that supports filesystem-based skills or prompt packs, as long as the folder is installed in the path that agent expects and SKILL.md remains the entrypoint.

---

🛠️ SAP Commerce Cloud CLI

This skill teaches AI agents how to operate the sapccm CLI tool to manage a CCV2 subscription on SAP Commerce Cloud Portal.

Highlights

• 🏗️ Create and monitor builds from any Git branch with live log streaming
• 🚀 Deploy builds to d1 / s1 / p1 environments with full control over update mode and deployment strategy
• 💾 Back up and restore data across environments with status polling
• 🔧 Inspect and update service properties (customer-properties, initial passwords, encryption keys)
• 🔐 Auth-failure recovery — prompts for updated client-id, client-secret, and token-url when credentials expire

Key Conventions Enforced

Reference Docs

• Commands — Full syntax cheat sheet for every subcommand
• Service Codes — All service codes and property codes with editability notes

Agent Compatibility

Security: The .env file in the skill directory holds your subscription code and is git-ignored. Never commit it. If authentication fails, update the CLI credentials via sapccm config set client-id / client-secret / token-url.

---

📚 Personal Knowledge Wiki

This skill teaches AI agents how to compile journals, notes, messages, and other personal data into a structured personal knowledge wiki.

Highlights

• 📝 Ingest source data from formats like Day One, Apple Notes, Obsidian, Notion, email, CSV, and plain text
• 🧠 Absorb entries into articles by turning raw entries into connected wiki pages instead of flat archives
• 🔎 Query the wiki to answer questions from synthesized knowledge rather than re-reading source files
• 🧹 Clean up and expand articles with audits, backlink rebuilding, and missing-page detection
• 🗂️ Organize knowledge into a durable wiki structure with index, backlinks, and absorb logs

Key Workflows

Agent Compatibility

This skill is compatible with any agent that supports filesystem-based skills or prompt packs, as long as the folder is installed in the path that agent expects and SKILL.md remains the entrypoint.

---

🎨 Excalidraw Diagrams

This skill teaches AI agents how to produce Excalidraw diagrams as clipboard JSON the user pastes straight onto the canvas. There is no canvas server, CLI, REST API, or MCP tool involved — the deliverable is a JSON file.

Highlights

• 🗂️ 17 diagram recipes — flowcharts, software architecture, sequence/UML/data-flow, mind maps, wireframes, Gantt charts, SWOT, lean canvas, personas, competitive matrices, brainstorm boards, presentation frames, and level/game maps
• 📐 Grid-planned layouts — tiers and x-positions decided before any JSON is emitted, everything snapped to a 20px grid
• 🎨 Design-system discipline — verified palette, matched stroke/fill pairs, max 4 accent colors each meaning one thing
• 🪣 Bucket fill regions — zone tinting, swimlane bands, and Gantt stripes via polygon line elements instead of stacked translucent rectangles
• ✅ Self-review pass — a 9-point static check (truncation, overlap, arrow geometry, crossings, spacing, reading order) before handover

Key Conventions Enforced

Reference Docs

• Design System — palette, typography, sizing, spacing, layering, anti-patterns
• Output Format — skeleton field reference, clipboard vs .excalidraw file, gotchas
• Diagram Recipes — per-type layout recipes and conventions for all 17 diagram types

Usage

Ask for a diagram in plain language — “draw the checkout flow”, “diagram this service architecture”, “make a mind map of these ideas”. The agent writes a JSON file, then you paste it:

cat .json | pbcopy

Click an empty spot on the Excalidraw canvas, press Cmd/Ctrl+V, then Shift+1 to zoom to fit.

Agent Compatibility

---

Repository Structure

.
├── .gitignore                             # Repository-level ignores
├── README.md                              # This file
├── sap-commerce-development/              # Skill package
│   ├── SKILL.md                           # Main skill definition
│   ├── LICENSE.txt                        # Apache 2.0 license
│   ├── agents/
│   │   └── openai.yaml                    # OpenAI agent config
│   └── reference/
│       ├── type-system.md
│       ├── patches.md
│       ├── spring-config.md
│       ├── occ-api.md
│       └── testing.md
├── sap-cloud-cli/                         # Skill package
│   ├── SKILL.md                           # Main skill definition
│   ├── LICENSE.txt                        # Apache 2.0 license
│   ├── .env                               # Local secrets — git-ignored
│   ├── agents/
│   │   └── openai.yaml                    # OpenAI agent config
│   └── reference/
│       ├── commands.md                    # Full command syntax cheat sheet
│       └── service-codes.md              # Service and property codes
├── wiki/                                  # Skill package
│   └── SKILL.md                           # Main skill definition
└── excalidraw-skill/                      # Skill package
    ├── SKILL.md                           # Main skill definition
    └── references/
        ├── design-system.md               # Palette, typography, sizing, layering
        ├── output-format.md               # Skeleton JSON field reference
        └── diagram-recipes.md             # 17 per-type layout recipes

Installation

Each skill can be installed individually. Copy or symlink the skill folder into the directory your agent uses for custom skills.

# Claude Code (personal/global)
mkdir -p ~/.claude/skills
cp -r /path/to/this-repo/sap-commerce-development ~/.claude/skills/
cp -r /path/to/this-repo/sap-cloud-cli ~/.claude/skills/
cp -r /path/to/this-repo/wiki ~/.claude/skills/
cp -r /path/to/this-repo/excalidraw-skill ~/.claude/skills/

# Claude Code (project-local)
mkdir -p .claude/skills
cp -r /path/to/this-repo/sap-commerce-development .claude/skills/
cp -r /path/to/this-repo/sap-cloud-cli .claude/skills/
cp -r /path/to/this-repo/wiki .claude/skills/
cp -r /path/to/this-repo/excalidraw-skill .claude/skills/

Note for sap-cloud-cli: After copying, populate .env in the skill directory with your credentials. The file is already git-ignored in this repository.

Or add the entire collection as a Git submodule:

git submodule add  agent-skills
ln -s agent-skills/sap-commerce-development .claude/skills/sap-commerce-development
ln -s agent-skills/wiki .claude/skills/wiki
ln -s agent-skills/excalidraw-skill .claude/skills/excalidraw-skill

Contributing

Contributions are welcome — whether improving existing skills or adding entirely new ones.

Improving an Existing Skill

1. Fork the repository
2. Edit the relevant files inside the skill’s directory
3. Test by using the skill with your AI assistant on a real project
4. Submit a pull request

Adding a New Skill

1. Create a new folder under the root directory with a descriptive name
2. Add a SKILL.md with YAML frontmatter (name, description, license) and instructions
3. Add a LICENSE.txt for the skill
4. Optionally add agent-specific configs in an agents/ subdirectory
5. Update this root README.md to list the new skill in the Available Skills table

Guidelines

• Keep instructions concise and example-driven
• Use generic/anonymized names in code samples
• Include both the what and the why for every rule
• Each skill should be fully self-contained

License

Individual skills may have their own licenses and metadata inside each skill directory. The SAP Commerce Cloud Development skill is licensed under the Apache License 2.0.