claude

workbench-dev-team

The Index-driven dev pipeline: Inspector Lestrade (triage), Dr. Watson (dev work), and Sherlock Holmes (code review) as subagents. Interactive or unattended dispatch.

0 on GitHubby Mike Bronnerv0.59.1updated yesterdaydevelopment
/

Plugin workbench-dev-team from marketplace claude-workbench in mike-bronner/claude-workbench. Commands read from the repo's manifest — today.

This plugin runs code on your machine — its hooks fire automatically on events like tool use. The exact commands are listed below; read them before installing.

What it installs

Skills (4)

  • comms-style

    How Lestrade, Watson, and Holmes write every piece of prose that isn't code — ticket comments, PR/issue descriptions, review verdicts, escalation notes, AC checklists. Modeled on ASD-STE100 (Simplified Technical English). This is a standing way of writing, not a checklist to run after drafting — app

  • develop

    Apply universal development standards when implementing code changes, fixing bugs, refactoring, writing tests, or any task that writes or modifies source code. Use this skill BEFORE writing or changing code — every time, manual or agent-driven — to ensure consistent conventions, testing, commit hygi

  • git-commit

    Generate commit messages using Conventional Commits + Gitmoji format. Use this skill whenever creating, drafting, or suggesting git commit messages — including /commit commands, pre-commit hooks, bulk commits, and any context where a commit message is being composed. Always invoke this skill before

  • orchestrate

    Run the dev team (Lestrade, Watson, Holmes) as background sub-agents. Holds the agent routing, the six-slot brief every handoff uses, and GitHub action routing. Read it before picking any sub-agent.

Slash commands (1)

  • /workbench-dev-team:setup

    Configure the workbench-dev-team plugin — verify prerequisites, check Keychain credentials, register The Index MCP, move an old agent config into /config, install the dispatcher, and install the Dispatch prompt into the workbench-dev-team-dispatch scheduled task. Re-run after a plugin update, after

Subagents (10)

  • holmes-index

    Sherlock Holmes in The Index mode only — the item-ID token in, one App-signed verdict on the item's PR out. Dispatch workbench-dev-team:holmes, which routes the token here.

  • holmes-lens

    Read-only helper that Sherlock Holmes dispatches inside his own review — a lens reviewer, a skeptic, a red-team attacker, a blue-team defender, or an auditor. It reads the tree it is pointed at, runs read-only commands and the existing tests that cover the change in that tree, and reports findings a

  • holmes-local

    Sherlock Holmes in Local mode only — a six-slot brief in, a prose verdict on the uncommitted working tree or the committed range (and PR number) its Workdir: names out, with no Index call and no GitHub write. Dispatch workbench-dev-team:holmes, which routes a brief here.

  • holmes

    Code review agent with two modes. Local mode is the default — any prose brief reviews the uncommitted working tree in the given workdir, or the committed range (such as origin/main...HEAD) its Workdir names, against the brief's Acceptance list as its rubric, with no The Index calls and no GitHub wri

  • lestrade-item

    Inspector Lestrade in Item mode only — the item-ID token in, acceptance criteria and WSJF scores on one item out. Dispatch workbench-dev-team:lestrade, which routes the token here.

  • lestrade-sweep

    Inspector Lestrade in Sweep mode only — the repo-sweep token in, blocked-by links and consolidated follow-ups across one repo out. Dispatch workbench-dev-team:lestrade, which routes the token here.

  • lestrade

    Triage agent. Two operating modes detected from input shape — Item mode (dispatched by Dispatch on one unrefined GitHub project item; inspects the issue + repo, generates acceptance criteria checked by a blind multi-lens fan-out before it's written, scores WSJF fields, moves the item to Backlog) and

  • watson-direct

    Dr. Watson in Direct mode only — a six-slot brief in, an uncommitted working tree and a proposed commit message out. Dispatch workbench-dev-team:watson, which routes a brief here.

  • watson-index

    Dr. Watson in The Index mode only — the item-ID token from bin/dispatch-agent.sh in, a ready PR and an In Review item out. Dispatch workbench-dev-team:watson, which routes the token here.

  • watson

    Development agent. Direct mode is the default — any prose brief runs the universal dev workflow with no The Index calls, for ad-hoc dev work delegated from Claude Code or Cowork. The Index mode is entered only on an explicit item-ID token, and runs the full pipeline orchestration: claim the item, fe

Hooks (4)

  • PermissionRequest

    bash "${CLAUDE_PLUGIN_ROOT}/hooks/scripts/pipeline-scope.sh"

  • SessionStart

    command -v node >/dev/null 2>&1 && node "${CLAUDE_PLUGIN_ROOT}/hooks/fallback/session-start.ts" || true

  • SubagentStart

    command -v node >/dev/null 2>&1 && node "${CLAUDE_PLUGIN_ROOT}/hooks/fallback/subagent-start.ts" || true

  • SubagentStop

    command -v node >/dev/null 2>&1 && node "${CLAUDE_PLUGIN_ROOT}/hooks/fallback/subagent-stop.ts" || true

View source on GitHub

More from mike-bronner/claude-workbench

All 5 plugins →

Core infrastructure: persistent agent identity, session lifecycle hooks, operational memory, and meta skills. Customizable agent name and memory path.

15 skills · 2 agents · 1 MCP server · hooks

0

Active write-back, human-approved, tax-aware YNAB budget review and categorisation assistant.

3 skills · 13 commands · 1 agent · 1 MCP server · hooks

0

BuJo ritual system: daily, weekly, monthly, and yearly interactive reflections in Apple Notes. Scheduled rituals with on-demand slash commands.

11 commands · 1 agent · 1 MCP server · hooks

0

Similar plugins

superpowers

claudecodex

Core skills library for Claude Code: TDD, debugging, collaboration patterns, and proven techniques

15 skills · hooks

obra/superpowers 316,836

context7

claudecodex

Up-to-date documentation lookup. Pull version-specific documentation and code examples directly from source repositories into your coding agent's context.

1 MCP server

upstash/context7 64,709

Build powerful, stateful CLI interfaces for any GUI application using the cli-anything harness methodology.

5 commands

HKUDS/CLI-Anything 58,665

Automates browser interactions for web testing, form filling, screenshots, and data extraction

12 skills

vercel-labs/agent-browser 50,373

Use Alibaba Open Code Review from local Codex.

2 skills

alibaba/open-code-review 50,068

slidev

claude

Claude Code skill for creating and presenting Slidev slidedecks

1 skill

slidevjs/slidev 49,452

Maintain workbench-dev-team?

This page updates itself from your repo. Add the badge to your README, or get it Featured to pin it above the organic results with a followed link to your site.

claude & codex plugins5 plugins