| name | conductor-context |
| description | Loads Conductor project context automatically. Use when user asks about the project, starts tracks, implements features, or needs project info. Auto-loads conductor/ files as needed. |
Conductor Context Skill
When I activate
- User mentions "project", "our app", "what are we building"
- User runs /conductor:setup, /conductor:track, or /conductor:implement
- User asks about tech stack, workflow, or style guides
- Need context for generating specs or plans
What I do
1. Check if Conductor is initialized
First, verify the project has Conductor set up:
test -d conductor/ && echo "initialized" || echo "not initialized"
If not initialized:
- Tell user: "Run
/conductor:setup first to initialize Conductor"
- STOP here
2. Load only what's needed for current task
For planning/new tracks:
- Read
conductor/product.md (2-3 sentence summary only)
- Read
conductor/tech-stack.md (list technologies only)
For implementation:
- Read
conductor/workflow.md (full content)
- Read active track from
conductor/tracks.md
- Read
conductor/tracks/<id>/spec.md
- Read
conductor/tracks/<id>/plan.md
For style questions:
- List available guides in
conductor/styleguides/
- Read specific guide only if asked
3. Provide concise summary
Format:
Project: [1 sentence from product.md]
Stack: [key technologies]
Active Track: [ID and title, if any]
Current Task: [Phase X, Task Y, if implementing]
Token Efficiency Rules
- Never load all files at once
- Never reproduce entire file contents
- Always summarize briefly (2-3 sentences max)
- Only load what's relevant to current conversation
- If brownfield: Read only package.json, README.md, and file tree
Important
This skill caches context in the conversation. Don't reload unless files change.