Claude-skill-registry-data mcp-development-brief-hq-brief-wiggum
Developing and testing MCP tools for Brief
install
source · Clone the upstream repo
git clone https://github.com/majiayu000/claude-skill-registry-data
Claude Code · Install into ~/.claude/skills/
T=$(mktemp -d) && git clone --depth=1 https://github.com/majiayu000/claude-skill-registry-data "$T" && mkdir -p ~/.claude/skills && cp -r "$T/data/mcp-development-brief-hq-brief-wiggum" ~/.claude/skills/majiayu000-claude-skill-registry-data-mcp-development-brief-hq-brief-wiggum && rm -rf "$T"
manifest:
data/mcp-development-brief-hq-brief-wiggum/SKILL.mdsource content
MCP Tool Development
Adding New MCP Tools
- Define tool in
mcp-server/src/tools/[domain]/[tool-name].ts - Add to registry in
mcp-server/src/tools/index.ts - Update types in
mcp-server/src/types.ts - Test with Claude Code MCP inspector
- Document in
.brief/brief-guidelines.md - Add to allowlist in
if needs preapproval.claude/settings.local.json
Tool Structure
export const myTool: Tool = { name: 'brief_my_operation', description: 'Clear description of what it does', inputSchema: { type: 'object', properties: { // Zod-like schema }, required: ['field1'] }, handler: async (args, context) => { // Implementation return { success: true, data: result }; } };
Testing MCP Tools
- Use MCP inspector:
npx @modelcontextprotocol/inspector - Test error cases (missing params, invalid auth, not found)
- Validate against OpenAPI spec if API-backed
- Test from Claude Code with actual workflows
Documentation Pattern
Update
.brief/brief-guidelines.md:
### New Operation Category - `operation_name` - Description of operation - Required params: param1, param2 - Optional params: param3 - Returns: what it returns - Example: JSON example
Common Mistakes
- ❌ Not validating input schema
- ❌ Not handling authentication errors
- ❌ Not documenting in guidelines
- ❌ Not adding to allowlist if needed
- ✅ Test happy path AND error paths