claude

flow

Conversational, spec-driven development workflow: /flow:help, /flow:init, /flow:run, /flow:hunt, /flow:review, /flow:pr, /flow:validate, /flow:lint, /flow:ship + independent spec/diff review agents, always-on spec/commit guard hooks, and multi-repo work items that cannot close while a sibling slice is unbuilt.

1 on GitHubby liminalarcv1.9.0MITupdated 1 month ago
/

Plugin flow from marketplace flow-toolkit in liminalarc/flow-toolkit. Commands read from the repo's manifest — 1 month ago.

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 (5)

  • hunt

    Hunt new feature opportunities through a domain-grounded persona panel, fanning research out to one flow:flow-researcher agent per dimension. Invoke as /flow:hunt [--deep|focus area].

  • pr

    Spec-aware review of a PR or branch diff — /flow:pr [pr# | branch] [--spec|--quality|--tests]

  • review

    Structured, multi-lens audit of a project — docs, UX, marketing, or product. Reads current state, finds gaps, proposes concrete fixes. Invoke as /flow:review [--docs|--ux|--marketing|--product]; runs all four lenses when none is named.

  • run

    Implement specs, manage backlog, brainstorm — routes each path (implement / --add / --ideas / --clean / --condense) loading only its reference. Invoke as /flow:run [spec#|--ideas|--add|--clean|--condense|description].

  • validate

    Live UI/UX validation — drive the running app and score it against a rubric via the flow-ux-validator agent. UI lens = design-system conformance; UX lens = task completion + friction. The entry point into UI/UX review; the reusable agent is also wired into /flow:run's done-gate (1.15). Invoke as /fl

Slash commands (4)

  • /flow:help

    One-screen map of everything flow can do — /flow:help [specs|build|review|agents|hooks|ambient]

  • /flow:init

    Bootstrap or adopt a project: index + specs/ detail files, CLAUDE.md, config — /flow:init [concept|--adopt|--greenfield|--backend ado]

  • /flow:lint

    Audit CLAUDE.md hierarchy + spec index/detail integrity; migrate legacy specs, reshape a flat spec, or rename specs to <id>-<slug> — /flow:lint [--claude|--specs|--fix|--migrate [id]|--rename [id|--all]]

  • /flow:ship

    Cut a release — /flow:ship [--dry-run]

Subagents (7)

  • flow-implementer

    Build ONE flow task against its approved local acceptance criteria. Dispatched by /flow:run's build loop (one per task file), post-sign-off — never for planning or scope decisions. Writes code test-first and stops at the seam.

  • flow-pr-reviewer

    Audit ONE flow-pr dimension (spec fidelity, quality, or tests) of a PR/branch diff against its rubric and return prioritized, actionable findings. Read-only — it audits the diff and reports, never edits. Dispatched by the flow-pr skill, one per dimension, to fan the review out in parallel; the main

  • flow-researcher

    Research ONE flow-hunt dimension (persona-aware) and return prioritized, scored opportunity findings. Read-only — it researches and reports, never edits. Dispatched by the flow-hunt skill, one per research dimension, to fan the hunt out in parallel; the main thread synthesizes.

  • flow-reviewer

    Audit ONE flow-review lens (docs, UX, marketing, or product) against its rubric and return prioritized, actionable findings. Read-only — it audits and reports, never edits. Dispatched by the flow-review skill, one per lens, to fan the audit out in parallel; the main thread synthesizes.

  • flow-spec-reviewer

    Judge ONE draft spec against the INVEST/SMART rubric before any code is written and return a structured PASS/CONCERNS verdict. Read-only — it critiques, it never authors or edits. Dispatched by /flow:run at implement §2 (pre-sign-off) and offered on --add; advisory under checkpoint, blocking under a

  • flow-ux-validator

    Drive a RUNNING app and validate ONE lens (UI = design-system conformance, or UX = task completion + friction) against its rubric, returning a prioritized read-only assessment. The only flow agent that runs the app instead of reading code. Read-only re project files — it drives, scores, and reports;

  • flow-verifier

    Independently check ONE flow implementer's diff against that task's local acceptance criteria and return a structured PASS/FAIL verdict. Read-only — it judges, it never fixes. Dispatched by /flow:run after each implementer, before the diff is integrated; gates blocking under auto-build, advisory und

Hooks (4)

  • PostToolUse

    bash "${CLAUDE_PLUGIN_ROOT}/hooks/flow-spec-guard.sh"

  • PostToolUse

    bash "${CLAUDE_PLUGIN_ROOT}/hooks/flow-claude-guard.sh"

  • PreToolUse

    bash "${CLAUDE_PLUGIN_ROOT}/hooks/flow-commit-guard.sh"

  • SessionStart

    bash "${CLAUDE_PLUGIN_ROOT}/hooks/flow-session-brief.sh"

View source on GitHub

Similar plugins

superpowers

claudecodex

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

15 skills · hooks

obra/superpowers 316,836

Backend API design, GraphQL architecture, workflow orchestration with Temporal, and test-driven backend development

9 skills · 1 command · 8 agents

wshobson/agents 42,461

beads

claudecodex

AI-supervised issue tracker for coding workflows

1 skill · 1 agent · hooks

gastownhall/beads 28,698

MiniMax AI skills library for frontend, fullstack, Android, iOS, shader, GIF sticker, document, presentation, spreadsheet, and multimodal media workflows

18 skills

MiniMax-AI/skills 13,828

worktrunk

claudecodex

Worktrunk is a CLI for Git worktree management, designed for parallel AI agent workflows. This plugin provides configuration guidance (LLM commit messages, project hooks, worktree paths), automatic activity tracking (🤖/💬 indicators in `wt list` showing active Claude sessions), and a `/wt-diff` pane showing each worktree's changes.

2 skills · hooks

max-sixty/worktrunk 12,481

gsd-core

claude

GSD Core is a meta-prompting, context engineering, and spec-driven development system for AI coding agents.

72 skills · 72 commands · 64 agents · hooks

open-gsd/gsd-core 11,494

Maintain flow?

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 pluginslisted