context-loader
Scannednpx machina-cli add skill Ibrahim-3d/conductor-orchestrator-superpowers/context-loader --openclawContext Loader Skill
Efficiently load and manage project context for Conductor's context-driven development workflow.
Trigger Conditions
Use this skill when:
- Starting work on a new track or feature
- User mentions: "load context", "project context", "get context"
- Beginning
/conductor:implementworkflow - Need to understand project structure without reading all files
Token Optimization Protocol
1. Respect Ignore Files
Before scanning files, check for:
.claudeignore- Claude-specific ignores.gitignore- Standard git ignores
# Check for ignore files
ls -la .claudeignore .gitignore 2>/dev/null
2. Efficient File Discovery
Use git for tracked files:
git ls-files --exclude-standard -co | head -100
For directory structure:
git ls-files --exclude-standard -co | xargs -n 1 dirname | sort -u
3. Priority Files (Read First)
| Priority | File Type | Examples |
|---|---|---|
| 1 | Manifests | package.json, Cargo.toml, pyproject.toml |
| 2 | Conductor | conductor/product.md, conductor/tech-stack.md |
| 3 | Track | conductor/tracks/<id>/spec.md, plan.md |
| 4 | Config | tsconfig.json, .env.example |
4. Large File Handling
For files over 1MB:
- Read first 20 lines (header/imports)
- Read last 20 lines (exports/summary)
- Skip middle content
Enforcement Rules (MANDATORY)
- Check file size via Bash
ls -labefore reading-
500KB: Read first 20 + last 20 lines only
-
1MB: Skip entirely, log as "skipped: too large"
-
- Stop after Tier 1-3 files. Tier 4 (config files) only if task-specific.
- Never load completed tracks — only active track spec.md + plan.md.
- Maximum 15 files per context load. If more are needed, prioritize by tier.
Context Loading Workflow
1. Load CLAUDE.md (if exists)
2. Load conductor/product.md (project vision)
3. Load conductor/tech-stack.md (technical context)
4. Load conductor/tracks.md (completed work — prevents duplicate effort)
5. Load current track spec.md (requirements)
6. Load current track plan.md (tasks — check [x] vs [ ] status)
Evaluate-Loop Integration
This skill is used by multiple loop agents:
loop-planner— loads context before creating a planloop-executor— loads context before implementing tasksconductor-orchestrator— loads context to determine current loop step
Critical: Always load tracks.md and check plan.md task markers to prevent duplicate work across sessions.
Response Format
After loading context, summarize:
## Project Context Loaded
**Product**: [one-line summary]
**Tech Stack**: [key technologies]
**Current Track**: [track name/id]
**Active Phase**: [current phase]
**Pending Tasks**: [count of [ ] tasks]
**Completed Tasks**: [count of [x] tasks]
**Loop Step**: [current Evaluate-Loop step — Plan/Execute/Evaluate/Fix]
Source
git clone https://github.com/Ibrahim-3d/conductor-orchestrator-superpowers/blob/master/skills/context-loader/SKILL.mdView on GitHub Overview
Context-loader quickly loads essential project context for Conductor's context-driven workflow, helping you start tracks or implement features without pulling in every file. It prioritizes key docs, respects ignore rules, and enforces strict size limits to keep token usage efficient.
How This Skill Works
Context-loader first respects ignore files (.claudeignore and .gitignore) and uses git ls-files to discover tracked files. It then loads content in priority order (manifests, Conductor docs, then track-related files) while enforcing size rules: for files over 500KB it reads only the first 20 and last 20 lines, and any file over 1MB is skipped. It stops after loading Tier 1-3 files; Tier 4 configs are loaded only if the task requires them. It always includes the current track's spec.md and plan.md and loads tracks.md to surface completed work and prevent duplicates. After loading, it outputs a Project Context Loaded summary including product, tech stack, current track, active phase, pending and completed tasks, and the loop step.
When to Use It
- Starting a new track or feature to scope work quickly.
- Launching the /conductor:implement workflow to align tasks and assumptions.
- Needing project structure insights without reading the entire repository.
- Preventing duplicate work by inspecting completed tracks via tracks.md.
- During planning or execution loops to ensure current context is up-to-date.
Quick Start
- Step 1: Initiate context-load at the start of work or when starting /conductor:implement.
- Step 2: The loader pulls CLAUDE.md (if present), conductor/product.md, conductor/tech-stack.md, conductor/tracks.md, and the active track's spec.md plus plan.md (only).
- Step 3: Review the Project Context Loaded summary to plan next actions and avoid duplicate effort.
Best Practices
- Always respect ignore files (.claudeignore and .gitignore) before scanning.
- Follow the Tier prioritization: Manifests, Conductor docs, Track files, then Config if needed.
- Limit loaded files to a maximum of 15; skip or defer very large files (>1MB).
- Never load completed tracks; only load the active track's spec.md and plan.md.
- Check tracks.md and plan.md markers to prevent duplicating work across sessions.
Example Use Cases
- Starting a new feature track: load product.md, tech-stack.md, and the active track's spec.md to outline tasks.
- Implementing a fix on an active track: load the current spec.md and plan.md while ignoring large history files.
- Onboarding a new engineer: surface project structure and key docs to ramp quickly.
- Quarterly project audit: load tracks.md to verify completed work and avoid rework.
- Before /conductor:implement: fetch current context to align requirements and technical decisions.
