notque/vexjoy-agent / skills/meta/toolkit
toolkit
VerifiedToolkit management: create and evaluate skills and agents, manage routing tables, generate Claude.md.
references/routing-table-updater/routing-format.md
View on GitHub# Routing Entry Format Specification
Routing metadata lives in two layers:
1. **Source of truth**: the YAML frontmatter `routing:` block in each `skills/*/SKILL.md` and `agents/*.md` file. Hand-authored, tracked in git.
2. **Generated artifact**: `skills/INDEX.json` and `agents/INDEX.json` (schema v2.0). Built from frontmatter by the repo generator scripts. Gitignored — regenerate to repair; the generators own ordering and formatting.
## Skill Frontmatter Routing Block
```yaml
---
name: routing-table-updater
description: "Maintain /do routing tables when skills or agents change."
user-invocable: false
routing:
triggers:
- "update routing tables"
- "routing drift"
category: meta-tooling
pairs_with:
- toolkit-evolution
---
```
| Field | Required | Purpose |
|-------|----------|---------|
| `name` | yes | Skill identifier; index key |
| `description` | yes | Intent text the router reads |
| `routing.triggers` | yes | Phrases that route to this skill |
| `routing.category` | yes | Coverage grouping |
| `routing.pairs_with` | no | Components commonly co-dispatched |
| `routing.not_for` | no | Negative routing examples |
| `routing.force_route` | no | High-confidence route on a single trigger match |
| `user-invocable` | no | `false` hides the skill from direct user invocation |
## Agent Frontmatter Routing Block
Same shape; agents add `complexity` (e.g., `Medium`, `Complex`, `Medium-Complex`) and use `description` as the router's short description.
## skills/INDEX.json Entry Shape
```json
{
"version": "2.0",
"generated": "2026-07-01T23:44:02Z",
"generated_by": "scripts/generate-skill-index.py",
"skills": {
"routing-table-updater": {
"file": "skills/meta/toolkit/references/routing-table-updater.md",
"description": "Maintain /do routing tables when skills or agents change.",
"triggers": ["update routing tables", "routing drift"],
"category": "meta-tooling",
"user_invocable": false,
"pairs_with": ["toolkit-evolution", "generate-claudemd"]
}
}
}
```
Optional per-entry fields when present in frontmatter: `not_for`, `force_route`, `agent`, `version`.
## agents/INDEX.json Entry Shape
The repo ships no agents since plan 009, so `agents/INDEX.json` is normally empty; private overlay agents use the same entry shape in `agents/INDEX.local.json`.
```json
{
"agents": {
"my-overlay-engineer": {
"file": "private-agents/my-overlay-engineer.md",
"short_description": "Ansible automation: playbooks, roles, collections, Molecule testing, Vault security",
"triggers": ["ansible", "playbook"],
"pairs_with": ["verification-before-completion"],
"complexity": "Medium-Complex",
"category": "infrastructure"
}
}
}
```
## Regeneration
```bash
cd $HOME/vexjoy-agent
python3 scripts/generate-skill-index.py # rebuilds skills/INDEX.json
python3 scripts/generate-agent-index.py # rebuilds agents/INDEX.json
```
PostToolUse hooks (`hooks/posttooluse-sync-skill-index.py`, `hooks/posttooluse-sync-agent-index.py`) run these automatically when a SKILL.md or agent file is written or edited. Manual regeneration covers bulk changes, deletes outside the harness, and corrupted index files.
## Validity Rules
- Every entry's `file` path exists on disk (zero phantom entries).
- Every on-disk skill/agent with valid frontmatter appears in its index.
- Triggers are specific phrases, unique across components where practical; resolve overlaps per `conflict-resolution.md`.
- Edit routing metadata in the source frontmatter, then regenerate — the index rebuild discards direct index edits.
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis. | Stars: 471k | Updated | Claude CodeCodex | Verified | Install |
| diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness. | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependencies | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| executing-plansobra/superpowersUse when executing an implementation plan in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is available | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |