By HatCloud
Spec-driven task workflow for Claude Code: a 6-phase lifecycle (Init/Design/Plan/Execute/Test/End) with plugin hooks, code/design/plan review, optional Linear and Telegram integration, and TDD discipline.
Reviews code changes (a diff against a plan) and reports issues by severity. Read-only — never edits code. Dispatched by the task workflow at per-task and full-review points; context (diff range, plan, design, dimension) is injected by the caller. Do NOT use standalone.
Reviews a design document (design.md) for structural soundness, coverage, and adversarial edge cases. Read-only — never edits files. Dispatched by the task workflow during Phase 2 design review rounds; context (design.md, prompt.md, round focus) is injected by the caller. Do NOT use standalone.
Reviews an implementation plan (plan.md) for fidelity to the design, executability, and risk. Read-only — never edits files. Dispatched by the task workflow at Phase 3 (plan fidelity review); context (plan.md, design.md) is injected by the caller. Do NOT use standalone.
Executes a single plan task (implements code/edits, runs verification) following injected steps and guardrails. Has write tools. Dispatched by the task workflow Phase 4 in auto/parallel-agents execution mode (independent batches). Do NOT use standalone.
[hat-flow bundled dep — invoked explicitly by the task workflow, not auto-triggered] Use when facing 2+ independent tasks that can be worked on without shared state or sequential dependencies
[hat-flow bundled dep — invoked explicitly by the task workflow, not auto-triggered] Use when receiving code review feedback, before implementing suggestions, especially if feedback seems unclear or technically questionable - requires technical rigor and verification, not performative agreement or blind implementation
[hat-flow bundled dep — invoked explicitly by the task workflow, not auto-triggered] Use when encountering any bug, test failure, or unexpected behavior, before proposing fixes
[hat-flow bundled dep — invoked explicitly by the task workflow, not auto-triggered] Use when about to claim work is complete, fixed, or passing, before committing or creating PRs - requires running verification commands and confirming output before making any success claims; evidence before assertions always
Use when dispatched by task skill to review design, plan, or code. Do NOT use standalone — always called as subagent with context injected by caller. 触发词: "review 设计", "review 计划", "review 代码", "审查设计", "审查计划"
Requires secrets
Needs API keys or credentials to function
Uses power tools
Uses Bash, Write, or Edit tools
Own this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimOwn this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimBased on adoption, maintenance, documentation, and repository signals. Not a security audit or endorsement.
This plugin requires configuration values that are prompted when the plugin is enabled. Sensitive values are stored in your system keychain.
linear_api_keyAPI key for the optional Linear integration. Leave empty to run without Linear (the linear plugin degrades gracefully).
${user_config.linear_api_key}English · 中文
Spec-driven task workflow for Claude Code — a disciplined 6-phase lifecycle (Init → Design → Plan → Execute → Test → End) with a plugin hook system, independent code/design/plan review, optional Linear and Telegram integration, and TDD discipline. It turns an ad-hoc coding request into a repeatable, resumable, reviewable engineering process.
phases.md.full / standard / lite / hotfix).phases.md to know
exactly where to pick a task back up.These tools must be on your PATH before installing:
jq — required. The plugin hook engine parses manifests/config with it;
without jq all task-workflow plugins (review/linear/git/timing) silently
fail to fire. Install: brew install jq / apt-get install jq.python3 (3.8+) — required. Several workflow helpers and bin scripts run
on it.node (with npx) — only needed for the optional Linear integration
(@hatcloud/linear-mcp is launched via npx).Running /task-setup performs a preflight check and reports anything missing.
/plugin marketplace add HatCloud/hat-flow
/plugin install hat-flow@hat-flow
No GitHub SSH key? The
owner/reposhorthand clones over SSH and fails withPermission denied (publickey)if you have no SSH key configured. This repo is public — use the explicit HTTPS URL instead:/plugin marketplace add https://github.com/HatCloud/hat-flow.git(If that still hits an SSH error, your git has aninsteadOfrule rewriting GitHub HTTPS→SSH; configure an SSH key or remove that rewrite.)
Then run /task-setup to configure Linear identity, optional Telegram
notifications, enabled plugins, and output language.
/task-setup — one-time first-run configuration (dependency preflight,
Linear, plugins, language)./task <your task description> — start a task; the orchestrator drives it
through all six phases./task (no args) in a later session — resume the in-progress task where it
left off./task-end — close out a finished task (final report, retrospective,
archive).| Plugin | What it does | Default |
|---|---|---|
review | Independent code/design/plan review at phase boundaries | on |
linear | Sync the task to a Linear issue (status + design/plan/archive comments) | auto (on when configured) |
git | Conventional commits, branch naming, dirty-tree checks | on |
tdd | Red-green-refactor enforcement per execute task | preset-dependent |
retrospective | Post-archive process review | preset-dependent |
Presets pick sensible defaults; /task-setup and per-task tuning let you
override any of them.
linear_api_key user config; the @hatcloud/linear-mcp
server is launched via npx. Without a key, the Linear plugin disables itself.telegram@claude-plugins-official plugin and run /telegram:configure.Bundles four adapted skills from obra/superpowers
(MIT) under a hatflow- prefix (hatflow-systematic-debugging,
hatflow-verification-before-completion, hatflow-dispatching-parallel-agents,
hatflow-receiving-code-review). To use the auto-triggering upstream versions,
install obra/superpowers directly — the hatflow- prefix keeps both side by side.
MIT — see LICENSE.
npx claudepluginhub hatcloud/hat-flow --plugin hat-flowComprehensive feature development workflow with specialized agents for codebase exploration, architecture design, and quality review
A growing collection of Claude-compatible academic workflow bundles. Covers scientific figures, manuscript writing and polishing, reviewer assessment, citation retrieval, data availability, paper reading, literature search, response letters, paper-to-PPTX conversion, and evidence-grounded Chinese invention patent drafting. Rules are organized as reusable skill folders with explicit workflows and quality checks.
Harness-native ECC operator layer - 67 agents, 271 skills, 92 legacy command shims, reusable hooks, rules, selective install profiles, and production-ready workflows for Claude Code, Codex, OpenCode, Cursor, and related agent harnesses
Upstash Context7 MCP server for up-to-date documentation lookup. Pull version-specific documentation and code examples directly from source repositories into your LLM context.
Comprehensive skill pack with 66 specialized skills for full-stack developers: 12 language experts (Python, TypeScript, Go, Rust, C++, Swift, Kotlin, C#, PHP, Java, SQL, JavaScript), 10 backend frameworks, 6 frontend/mobile, plus infrastructure, DevOps, security, and testing. Features progressive disclosure architecture for 50% faster loading.
Tools to maintain and improve CLAUDE.md files - audit quality, capture session learnings, and keep project memory current.