A universal repository of AI agent configurations (skills, rules, and commands) with an installer that works across multiple AI-assisted IDEs.
| IDE | Skills | Rules | Commands |
|---|---|---|---|
| Cursor | ~/.cursor/skills/ |
~/.cursor/rules/ |
~/.cursor/agents/ |
| OpenCode | Converted to agents | AGENTS.md |
~/.config/opencode/agents/ |
| VSCode/Copilot | N/A | copilot-instructions.md |
N/A |
# One-liner install (interactive)
curl -fsSL https://raw.githubusercontent.com/testified/agent-skills/main/install.sh | bash
# Or clone and run locally
git clone https://github.com/testified/agent-skills.git
cd agent-skills
./install.shJust run the script with no arguments:
./install.shYou'll be prompted to:
- Select which IDEs to install to
- Choose symlink or copy mode
- Select which components (skills, rules, commands)
# Install to Cursor with symlinks
./install.sh --cursor --symlink
# Install to all IDEs with copies
./install.sh --all --copy
# Install only rules to OpenCode
./install.sh --opencode --rules-only| Option | Description |
|---|---|
--cursor |
Install to Cursor |
--opencode |
Install to OpenCode |
--vscode |
Install to VSCode/Copilot |
--all |
Install to all detected IDEs |
--symlink |
Use symlinks (auto-updates from repo) |
--copy |
Copy files (static snapshot) |
--skills-only |
Only install skills |
--rules-only |
Only install rules |
--commands-only |
Only install commands |
--no-backup |
Skip backup of existing configs |
-h, --help |
Show help message |
- Changes in this repo automatically reflect in all IDEs
- Update everywhere with
git pull - Single source of truth
- Requires repo to stay in place
- Static snapshot of configurations
- No dependency on repo location
- Re-run script to apply updates
- Good for distributing to others
Reusable agent capabilities:
| Skill | Description |
|---|---|
api-testing |
Design and review API tests including contract testing, schema validation, error handling verification, and performance considerations |
conventional-commits |
Format git commits using Conventional Commits standard |
code-review |
Systematic code review following best practices |
test-planning |
Create comprehensive test plans with scope, resources, risks, and criteria |
exploratory-testing |
Guide exploratory testing sessions including charter creation, session-based testing, heuristics application, and finding documentation |
test-case-design |
Design comprehensive test cases using boundary analysis, equivalence partitioning, decision tables, and state transition testing |
incident-triage |
Triage production incidents with severity assessment, root cause analysis, and escalation |
Coding standards and conventions:
| Rule | Description |
|---|---|
code-quality.mdc |
Naming conventions, functions, error handling |
pull-request-template.mdc |
Standard PR format with summary, changes, issues |
quality-metrics.mdc |
Standard quality metrics with recommended thresholds |
ci-cd-quality-gates.mdc |
Quality gates for CI/CD pipelines including coverage, security, performance |
test-data-management.mdc |
Test data isolation, cleanup, sensitive data handling, environment configs |
test-environment-checklist.mdc |
Checklists for test environment setup, validation, and maintenance |
Custom subagent definitions:
| Command | Description |
|---|---|
code-reviewer.md |
Specialized code review agent |
regression-impact.md |
Analyzes code changes to recommend regression test scope and prioritization |
meeting-notes-evaluator.md |
Evaluates meeting notes to extract test info, todos, and risks |
defect-analysis.md |
Analyzes defect patterns with root cause categorization, trend analysis, and hotspot identification |
agent-skills/
├── README.md # This file
├── install.sh # Universal installer
├── skills/ # Agent skills
│ ├── api-testing/
│ │ └── SKILL.md
│ ├── code-review/
│ │ └── SKILL.md
│ ├── conventional-commits/
│ │ └── SKILL.md
│ └── test-planning/
│ ├── exploratory-testing/
│ │ └── SKILL.md
│ └── test-case-design/
│ ├── code-review/
│ │ └── SKILL.md
│ └── incident-triage/
│ └── SKILL.md
├── rules/ # Coding standards
│ ├── code-quality.mdc
│ └── pull-request-template.mdc
├── commands/ # Subagent definitions
│ └── code-reviewer.md
└── memory-bank/ # Project documentation
- Create a directory:
skills/your-skill-name/ - Add
SKILL.mdwith required frontmatter:
---
name: your-skill-name
description: Brief description for discovery
version: 1.0.0
---
# Your Skill Name
## When to Use
- Trigger conditions
## Instructions
- What the agent should do- Submit a pull request
- Create a file:
rules/your-rule.mdc - Add required frontmatter:
---
description: One-line description
alwaysApply: true
---
# Your Rule
Instructions...- Submit a pull request
- Create a file:
commands/your-command.md - Add required frontmatter:
---
name: your-command
description: What this agent does
mode: subagent
model: fast
tools:
- read
- write
---
# Your Command
## System Prompt
Instructions for the subagent...- Submit a pull request
cd /path/to/agent-skills
git pull
# Changes are automatically reflectedcd /path/to/agent-skills
git pull
./install.sh --cursor --copy # Re-run with same optionsThe installer automatically creates backups at:
~/.agent-skills-backup/
└── 2024-01-21T10-30-00/
├── cursor/
└── opencode/
To restore:
# Find your backup
ls ~/.agent-skills-backup/
# Copy back manually
cp -r ~/.agent-skills-backup/[timestamp]/cursor/skills/* ~/.cursor/skills/Make sure you're running the script from the repo root:
cd /path/to/agent-skills
./install.sh- Verify the skill was installed:
ls ~/.cursor/skills/ - Restart Cursor
- Check for YAML frontmatter errors in SKILL.md
- Verify file exists:
cat ~/.config/opencode/AGENTS.md - Restart OpenCode
- Check file permissions
MIT