notque/vexjoy-agent / skills/meta/do
do
VerifiedClassify user requests and route to the correct skill, pipeline, and stack. Primary entry point for all delegated work.
| File | Kind | Size |
|---|---|---|
| SKILL.md | skill md | 31.2 KB |
| references/antigravity-execution.md | markdown | 21.2 KB |
| references/error-handling.md | markdown | 478 B |
| references/execution-architecture.md | markdown | 1.2 KB |
| references/hooks-guide.md | markdown | 2.7 KB |
| references/lazy-completion-detector.md | markdown | 2.2 KB |
| references/model-selection.md | markdown | 8.4 KB |
| references/parallel-analysis.md | markdown | 13.4 KB |
| references/perspective-prompts.md | markdown | 5.8 KB |
| references/pipeline-guide.md | markdown | 1.6 KB |
| references/planning-guide.md | markdown | 1.2 KB |
| references/progressive-depth.md | markdown | 5.4 KB |
| references/quality-gates.md | markdown | 536 B |
| references/quality-loop.md | markdown | 6.8 KB |
| references/repo-architecture.md | markdown | 4.9 KB |
| references/routing-telemetry.md | markdown | 4.6 KB |
| references/semantic-first-ab-results.md | markdown | 1.6 KB |
| references/workflow-dispatch.md | markdown | 9.5 KB |
| references/worktree-rules.md | markdown | 5.0 KB |
references/repo-architecture.md
View on GitHub# Repository Architecture
This repository contains skills, pipelines, hooks, and scripts for Claude Code. The repo ships no specialist agents (retired in plans/009): the dispatch worker is the built-in `general-purpose` agent and the skill carries the domain. Private overlay agents (`agents/INDEX.local.json`) still install and validate when named.
## Component Types
| Component | Location | Purpose | Format |
|-----------|----------|---------|--------|
| **Skill** | `skills/*/SKILL.md` | Domain (e.g., Go, Python, K8s) or methodology (e.g., TDD, debugging) | Markdown with YAML frontmatter |
| **Pipeline** | `skills/process/workflow/references/*.md` | Multi-phase structured workflow (same format as skills) | Markdown with YAML frontmatter |
| **Hook** | `hooks/*.py` | Event-driven automation | Python script |
| **Script** | `scripts/*.py`, `scripts/*.sh` | Deterministic operations | Python or shell script |
> **Note**: Pipelines are skills with explicit numbered phases and gates. They live in `skills/process/workflow/references/` for organizational clarity but are synced to `~/.claude/skills/` at install time, so Claude Code discovers them as regular skills.
## Key Frontmatter Fields
```yaml
---
name: skill-name
description: Brief purpose description
user-invocable: false # Hide from slash menu (internal skills)
context: fork # Run in isolated sub-agent context
model: opus # Optional; omit to inherit. opus | fable | inherit; haiku only with a read-only `tools:` list; never sonnet (`/do` SKILL.md, Model choice)
allowed-tools: # YAML list format
- Read
- Write
- Bash
hooks: # Skill-specific lifecycle hooks
PostToolUse:
- type: command
command: python3 -c "..."
timeout: 3000
routing: # Routing metadata
triggers: [keyword1, keyword2]
pairs_with: [skill1, skill2] # skills or pipelines only
complexity: Simple | Medium | Complex
category: domain | devops | meta
---
```
## Skill Reference Files
Skills use progressive disclosure. Domain knowledge lives in `references/`, loaded on demand based on task context. A retired agent's body lives at `<skill>/references/<slug>.md` and its references at `<skill>/references/<slug>/`.
### Structure
```
skills/frontend/frontend/
├── SKILL.md # Skill body with the Reference Loading Table
└── references/
├── react-native.md # Former agent body, loaded for React Native tasks
└── react-native/
├── list-performance.md # Loaded for list/scroll tasks
└── navigation-patterns.md # Loaded for navigation tasks
```
### Reference Loading Table (mandatory for skills with references)
Every skill with a `references/` directory MUST have a reference loading table in its SKILL.md. `build-dispatch.py` tells every worker to consult it. The table maps task keywords to specific reference files so the model knows WHICH file to load for the current task — not all of them.
```markdown
| Task involves | Load reference |
|---|---|
| Lists, FlatList, scrolling | `list-performance.md` |
| Animation, gestures, transitions | `animation-patterns.md` |
| Navigation, screens, routing | `navigation-patterns.md` |
```
Without this table, the worker either loads nothing (missing context) or everything (wasted tokens). Both defeat progressive disclosure.
### Reference File Standards
| Rule | Requirement |
|---|---|
| **Size** | <= 500 lines per file. Split at ~400 if growing. |
| **Framing** | Joy-checked: lead with what patterns ENABLE, not what they PREVENT. Use "Instead of:" / "Use:" not "Bad:" / "Good:" or "Incorrect:" / "Correct:". |
| **Structure** | Each section: `## Title` + `**Impact:** LEVEL` + explanation + code examples |
| **Isolation** | Reference only files in the same skill's directory. No cross-skill references. |
| **Comment header** | First line after title: `<!-- Loaded by {skill-name} when task involves {domain} -->` |
### Validation
Run `scripts/validate-references.py --all` to check:
- Every declared reference file exists on disk
- Every file on disk is declared in the SKILL.md body
- Structural compliance (headings, content, code examples)
Run `scripts/validate-references.py --check-size` to check reference size limits (500 lines) against the dated debt register and flag empty `references/` dirs.
Run `python3 -m pytest scripts/tests/test_reference_loading.py` to verify:
- Reference loading tables are complete
- Keyword→reference mapping selects the right files
### When Adding Domain Knowledge
Add it to the skill that owns the domain; do not create a repo agent.
1. Add reference files under `skills/<group>/<skill>/references/` (one per domain concern, <= 500 lines each)
2. Add rows to the skill's reference loading table mapping keywords → files
3. Regenerate `skills/INDEX.json` via `scripts/generate-skill-index.py` when frontmatter changes
4. Run `scripts/validate-references.py --all` to verify
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| writing-skillsobra/superpowersUse when creating new skills, editing existing skills, or verifying skills work before deployment | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| xlsxanthropics/skillsUse this skill any time a spreadsheet file is the primary input or output. This means any task where the user wants to: open, read, edit, or fix an existing .xlsx, .xlsm, .xltx, .csv, or .tsv file (e.g., adding columns, computing formulas, formatting, charting, cleaning messy data); create a new spreadsheet from scratch or from other data sources; or convert between tabular file formats. Trigger especially when the user references a spreadsheet file by name or path — even casually (like "the xlsx in my downloads") — and wants something done to it or produced from it. Also trigger for cleaning or restructuring messy tabular data files (malformed rows, misplaced headers, junk data) into proper spreadsheets. The deliverable must be a spreadsheet file. Do NOT trigger when the primary deliverable is a Word document, HTML report, standalone Python script, database pipeline, or Google Sheets API integration, even if tabular data is involved. | Stars: 198k | Updated | Claude CodeCodex | Verified | Install |
| shipping-and-launchaddyosmani/agent-skillsPrepares production launches. Use when preparing to deploy to production, or when asking what needs to be in place before shipping. Use when you need a pre-launch checklist, when setting up monitoring, when planning a staged rollout, or when you need a rollback strategy. | Stars: 119k | Updated | Claude CodeCodex | Verified | Install |
| childhoodthedotmack/claude-memBabysit a shipped release after merge, publication, or deployment. Verify published packages, production behavior, installed plugins, and late CI or user reports until the observation window is clean; use babysit for PR readiness before release. | Stars: 118k | Updated | Claude CodeCodex | Verified | Install |