Skip to content
Development
Command

/d

Jev-first router: A/B variant of /do. One TypeSafe call replaces the manifest read; falls back to /do when unavailable or unconfident.

From plugin
vexjoy-agent
42012 skills198 agents12 commands77 hooks
Install
$ npx -y skills add notque/vexjoy-agent --agent claude-code

How it fires

How this command gets triggered: by you, by Claude, or both.

  • Fires itselfClaude auto-loads it when your prompt matches the work.
  • You can call itInvoke it directly when you want it.
  • Slash command/d

Context preview

What this command does when you run it.

Jev-first router: A/B variant of /do. One TypeSafe call replaces the manifest read; falls back to /do when unavailable or unconfident.

Command definition

d.md
description: "Jev-first router: A/B variant of /do. One TypeSafe call replaces the manifest read; falls back to /do when unavailable or unconfident."
argument-hint: "[request]"
allowed-tools: ["Read", "Bash", "Grep", "Glob", "Skill", "Task"]

/d - Jev-first Router (A/B variant of /do)

Route user requests using a Jev classifier call instead of `/do`'s in-context manifest read. Use `/d` to A/B-test against `/do`; use `/do` for normal work.

Instructions

Call the Skill tool with `d`. Follow its classify, decide, enhance, and execute phases.

Base directory: $CLAUDE_PROJECT_DIR or the current working directory
Skill file: skills/meta/d/SKILL.md

**Phase 1: CLASSIFY** — One `scripts/jev-route.py` call: deterministic force-route guard, then (if not force-routed) the merged Jev classification. **Phase 1F: FALLBACK** — On `fallback: true` (TypeSafe unavailable, low confidence, or error), read `skills/meta/do/SKILL.md` in full and run its Phase 1-4 unmodified. **Phase 2: DECIDE** — Apply the Jev/force-route decision directly, display the routing banner. **Phase 3: ENHANCE** — Map stack signals to concrete stack entries. **Phase 4: EXECUTE** — Same Task Spec + `build-dispatch.py` contract `/do` uses, reused unmodified.

The routing banner MUST be the first visible output:

===================================================================
 ROUTING (/d): [brief summary]
===================================================================

 Selected:
   -> Agent: [name]
   -> Skill: [name]
   -> Source: [jev|pre-route-force] (confidence: [level])

 Invoking...
===================================================================

On fallback, the banner instead states `ROUTING (/d -> falling back to /do): [reason]` and control passes entirely to `/do`'s own instructions.

For the full design (why `/d` exists as a separate command, the request/response contract, fallback-condition reasoning, measured eval results), read `skills/meta/d/SKILL.md` and `skills/meta/d/references/jev-classifier-design.md`.

ARGUMENTS: $ARGUMENTS

Read more
Ships withvexjoy-agent

Essays and writing behind this toolkit live at vexjoy.com. VexJoy Agent connects plain-English requests to specialist agents, skills, and workflows. /do selects the knowledge and tools needed for your task.

Get the whole plugin

Other commands on vexjoy-agent.