Claude-skill-registry conductor-development
Context-Driven Development skill for projects using Conductor. Use this skill when you detect a `conductor/` directory in the project, when working on tasks defined in a `plan.md` file, or when the user asks about tracks, specs, or plans. Automatically applies TDD workflow, tracks task completion, and follows the spec-driven methodology.
git clone https://github.com/majiayu000/claude-skill-registry
T=$(mktemp -d) && git clone --depth=1 https://github.com/majiayu000/claude-skill-registry "$T" && mkdir -p ~/.claude/skills && cp -r "$T/skills/data/conductor-development" ~/.claude/skills/majiayu000-claude-skill-registry-conductor-development && rm -rf "$T"
skills/data/conductor-development/SKILL.mdConductor Development Skill
This skill enables Claude to work effectively on projects managed by the Conductor framework - a spec-driven, structured development methodology.
When This Skill Activates
Claude should automatically apply this skill when:
- A
directory exists in the project rootconductor/ - The user mentions "tracks", "conductor", "spec", or "plan" in the context of development
- Files like
,conductor/tracks.md
, orconductor/workflow.md
are presentconductor/product.md - The user runs any
command/conductor:*
Core Principles
- The Plan is the Source of Truth: All work must be tracked in
plan.md - Spec-Driven Development: Understand the spec before implementing
- Test-Driven Development: Write tests before implementation (Red-Green-Refactor)
- High Code Coverage: Target >80% code coverage for all modules
- Track Progress: Update task status markers (
→[ ]
→[~]
)[x]
Project Structure Understanding
When working on a Conductor project, familiarize yourself with:
conductor/ ├── product.md # Product vision and goals ├── product-guidelines.md # Brand voice and communication style ├── tech-stack.md # Technology choices and constraints ├── workflow.md # Development methodology and procedures ├── tracks.md # Master list of all tracks ├── code_styleguides/ # Language-specific style guides │ ├── general.md │ ├── python.md │ └── [others] └── tracks/ # Individual track folders └── <track_id>/ ├── spec.md # Feature specification ├── plan.md # Implementation plan with tasks └── metadata.json # Track metadata
Task Execution Protocol
When implementing a task from a Conductor plan:
1. Select and Mark Task
- Find the next pending task (marked
) in the track's[ ]plan.md - Update its status to in-progress:
[~]
2. Follow TDD Workflow
- Red Phase: Write failing tests that define expected behavior
- Green Phase: Write minimal code to make tests pass
- Refactor: Improve code while keeping tests green
3. Complete the Task
- Verify test coverage (>80%)
- Follow code style guidelines from
conductor/code_styleguides/ - Commit with semantic message (e.g.,
)feat(auth): Add login form - Mark task complete:
and append commit SHA[x] - Commit the plan update:
conductor(plan): Mark task 'X' as complete
4. Phase Completion
When completing a phase:
- Run full test suite
- Provide manual verification steps to user
- Create checkpoint commit with verification report
- Update
with checkpoint SHAplan.md
Status Markers
- Pending (not started)[ ]
- In Progress (currently working)[~]
- Completed (done with commit SHA)[x]
Available Commands
Remind users of available Conductor commands:
- Initialize Conductor in a project/conductor:setup
- Create a new feature/bug track/conductor:new-track
- Execute tasks from the current track/conductor:implement
- Show project progress/conductor:status
- Git-aware revert of tracks/phases/tasks/conductor:revert
Context Loading
Before starting any implementation work, always load:
- For task lifecycle proceduresconductor/workflow.md
- For technology constraintsconductor/tech-stack.md- The active track's
- For requirementsspec.md - The active track's
- For current task statusplan.md
Git Integration
Conductor tracks work through Git:
- Each task completion = one code commit + one plan commit
- Phase checkpoints create special commits with git notes
- SHAs are recorded in
for traceabilityplan.md - Use semantic commit messages following the project's conventions
Error Handling
If something goes wrong:
- Do not proceed without user confirmation
- Announce the failure clearly
- Suggest recovery options (e.g.,
)/conductor:revert - Wait for user instructions
Quality Gates
Before marking any task complete, verify:
- All tests pass
- Code coverage meets requirements (>80%)
- Code follows project's style guidelines
- Public functions are documented
- No linting errors
- No security vulnerabilities introduced