dev-workflow

Plugin dev-workflow from marketplace indie-toolkit in n0rvyn/indie-toolkit. Commands read from the repo's manifest — 4 days ago.

Skills (29)

• afk

  Use when the user is stepping away and wants work driven to a stated end state on its own — '/afk', 'afk', '我出去一趟', '你自己跑', '一直推到 block', '推到头', 'run this while I'm out', 'push it to the end', \"don't ask me, just decide\". Also use when a verified multi-phase dev-guide exists and needs driving acro

• audit-tokens

  Use when the user says 'audit tokens', 'token audit', 'analyze token consumption', '审计 token', '分析 token 消耗', '查 token 开销', 'cost audit', 'token consumption analysis', or wants a multi-dimensional analysis of Claude Code session token consumption with HTML report. Accepts integer days or natural-lan

• brainstorm

  Use before any creative work — creating features, building components, adding functionality, or modifying behavior. Explores user intent, requirements and design through iterative one-question-at-a-time conversation before implementation. Not when: task is mechanical with clear specs (use write-plan

• collect-lesson

  Use after resolving a bug, completing a task with notable pitfalls, or discovering an important architecture constraint. Extracts a structured entry from the session and saves it to the knowledge base. Not when: settling pre-implementation design decisions before /write-plan — use crystallize instea

• commit

  Use when the user says 'commit' or wants to save progress after completing a task. Analyzes uncommitted changes, groups them logically, and commits with conventional format messages.

• crystallize

  Use before /write-plan or when the session has settled significant design decisions. Extracts the decision history from the current conversation into a persistent crystal file that /write-plan and plan-verifier consume. Not when: extracting a postmortem after a fixed bug or completed task — use coll

• design-decision

  Use when facing multiple design or product options, the user is stuck between approaches, or asks '选哪个', '怎么设计', 'how should I design this', 'which is better'. Analyzes trade-offs by separating essential from accidental complexity and presents a structured comparison. Not when: options are not yet i

• distill-discussion

  Use when you have raw discussion files (AI conversation exports, exploratory notes) and want to extract structured outputs. Routes each file to: crystal (decisions), lesson (failures/pitfalls), or skip (no conclusion).

• execute-plan

  Use when the user says 'execute the plan', 'run the plan', 'implement the plan', '执行计划', '跑计划', or has a verified plan file ready for execution. Executes plan tasks in segments via the Workflow tool, with explicit hard-stop checkpoint gates (batch 1 always; any batch with a dependency hub; any batch

• finalize

  Use after all phases of a dev-guide are complete, or when the user says 'finalize', 'final check', 'cross-phase validation'. Runs full test suite, verifies acceptance criteria across all phases, audits cumulative test coverage, and produces a final validation report.

• finish-branch

  Use when implementation is complete and you need to decide how to integrate the work, or the user says 'finish branch', 'finalize branch', '完成分支', '结束分支', '收尾', 'merge or pr'. Runs full test suite, then presents structured integration options (merge to main / open PR / archive branch) and handles cl

• fix-bug

  Use when the user reports an error with stack trace or screenshot, describes unexpected behavior, build/test failures occur, OR provides a batch of issues to fix against a running system that exposes an end-to-end verification surface — API, CLI, REPL, chat agent, or mobile deeplink ('fix these N is

• fork-this

  Use when discussing topic A and an orthogonal problem B surfaces mid-session, OR the user says 'fork this', 'park this', '先记下来', '先放着', 'fork-this', '分叉', '现场分叉', '另起一个'. Generates a minimal seed prompt for B in a clean fork context, writes to docs/06-plans/ (or .claude/parked/ for non-dev-workflow

• generate-bases-views

  Generate Obsidian Bases (.base) views for structured querying of crystals, lessons, and vault notes. Use when the user says 'generate views', 'create base files', 'make queryable', '生成 base views', '生成视图', '让它可查询', or after crystallize/collect-lesson.

• generate-design-prompt

  Manual /generate-design-prompt invocation only (auto-routing disabled). Generates a design tool prompt (for Stitch, Figma, or similar) from actual project functionality — initial prompts from design docs/code or refinement prompts from design analysis feedback. Output adapts to detected platform (iO

• handoff

  Use when ending the current session: the user says 'handoff', '交接', 'handoff or done', or '收尾' to close the session. It decides between writing a full transfer doc for a new session (next day, different person) and closing with nothing to hand off, and in both cases screens the session for a lesson

• issue

  Use when the user says '/issue', 'file an issue', 'record this as an issue', or wants to list/read/create GitHub Issues. Unified GitHub Issue entry point — list and read are cheap; create may include hypothesis generation (judgment), so the skill inherits the session model rather than downgrading. N

• kb

  Use when the user says '/kb', 'search kb', 'find lesson about X', '搜知识库', '以前怎么处理 X', or needs to search the cross-project knowledge base manually. Accepts a query, searches ~/.claude/knowledge/ for past lessons, API gotchas, architecture decisions, and learning notes. Not when: user wants to ADD a

• next-increment

  Use when the user says 'next increment', 'next step', 'pick the next evolution step', '下一步做什么', '下一步', or asks what to build next on a mature codebase with established architecture docs. Generates 3-5 candidate increments grounded in the architecture, then writes a mini-spec for the chosen one. Not

• review-before-commit

  Use when the user says 'review changes', 'review-before-commit', '审查变更', '检查改动', '提交前审查', 'pre-commit review', or wants a semantic review of uncommitted changes before committing. Classifies changes into enhancements, fixes, refactors, and removals; explains what each does; detects breaking changes;

• review-execution

  The single review dispatcher for this marketplace. Use when the user says 'review execution', 'parallel review', 'deep review', 'review my code', 'review after coding', 'execution review', '审查执行', '并行 review', '写完 review 一下', '代码 review 一下', '深度代码审查', '执行后审查', or wants a fresh-context multi-lens rev

• run-phase

  Use when the user says 'run phase', 'start phase N', 'next phase', '继续开发', '跑下一阶段', '开始第N阶段', or when continuing development guided by a dev-guide. Orchestrates the plan-execute-review cycle for one phase of a development guide: write-plan → verify-plan → execute-plan → test-changes → review agents

• self-pacing

  Moved: use /afk (dev-guide mode) to drive a verified dev-guide across phases; for one plan use /execute-plan. Triggers: '/self-pacing', 'self-pace', '跨阶段连跑', '不要每个阶段都停'.

• test-changes

  Use when the user says 'test changes', 'run tests', 'test the build', or after execute-plan completes in a run-phase. Runs the project's build/test/lint suite and returns a filtered report with errors only. Branches on project type: Apple projects (.xcodeproj/.xcworkspace) run xcodebuild from the ma

• understand-design

  Use when the user has a design prototype (from Stitch, Figma, or any tool) and wants AI to analyze it. Works standalone for daily development or in pipeline mode with a design doc. Handles images, HTML/CSS, or both. Not when: user wants the prototype audited AGAINST an existing implementation (use a

• verify-plan

  Use when a plan has been written and needs validation before execution, or the user says 'verify plan', 'check the plan', 'review the plan', 'validate plan', '检查计划', '验证计划'. VERIFY-ONLY — does not execute the plan; invoke execute-plan separately. Applies Verification-First method with falsifiable er

• write-dev-guide

  Use when starting a new project's development after design is approved, OR when scoping a multi-unit refactor/migration that spans multiple independent components (e.g., 'refactor 6 cards across 4 tabs', 'migrate auth across 4 layers'), or the user says 'write dev guide', 'break down this project in

• write-feature-spec

  Generate feature spec (docs/05-features/) capturing what was built, design decisions made, and user journeys completed. Use for cross-session context recovery, publication notes, or as input for external review. Not a prerequisite for self-review.

• write-plan

  Use when the user says 'write a plan', 'plan this', 'break this into tasks', '写计划', '拆分任务', or has requirements/specs for a multi-step task before touching code. Also invoked by `fix-bug` for large fixes (multi-file, architectural, or replacing a design), which pass their diagnosis evidence into the

Subagents (11)

• bug-diagnoser

  Read-only root-cause diagnosis. Turns a described failure (error text, stack trace, log, failing test, blocked task) into a confirmed cause with file:line evidence, every other site with the same cause, the affected consumers, and a bug-vs-design verdict. When given earlier failed hypotheses, it tes

• change-classifier

  Single-pass semantic review of uncommitted changes before a commit: applies the mechanical 200-line depth rule, classifies change groups as enhancement / fix / refactor / removal, detects breaking changes (Lens C checklist + residual-caller grep), compares against a plan's Impact Map when a plan pat

• design-analyzer

  Use this agent to analyze design prototypes from Stitch, Figma, or other design tools. Performs dual-channel analysis (image + code simultaneously), extracts design tokens, translates to target platform, and validates against UX assertions when in pipeline mode. Examples: <example> Context: User cre

• dev-guide-verifier

  Use this agent to verify a development guide after write-dev-guide produces it and before presenting to the user. Checks feature coverage, dependency graph semantics, cross-phase data flow, existing code overlap, term definitions, acceptance criteria quality, and structural integrity. Examples: <exa

• distill-discussion-reader

  Reads a single discussion file and classifies it by conclusion type. Returns structured extraction (crystal fields, lesson fields, or skip reason). Used by the distill-discussion skill for large files (>100KB) that need context isolation. Examples: <example> Context: Large discussion file needs clas

• execute-plan

  Use this agent to execute a single task (or a small batch) of a verified implementation plan. Invoked by the execute-plan Workflow script via agent() with model=sonnet. Receives a canonical task id ("<N>" / "<N>-tests" / "<N>-impl") and a checkpoint-file path, runs the task's **Verify:**, and return

• feature-spec-writer

  Use this agent to generate feature specs comparing design intent against implementation. Produces docs/05-features/ documents for use by the apple-dev:feature-reviewer agent (auto-dispatched by run-phase via /review-execution). Examples: <example> Context: A feature is complete and needs documentati

• flow-tracer

  Use this agent to trace a call chain end-to-end through the codebase given a natural language flow description. Identifies each hop with file:line and reports breaks (signal with no consumer, call to nonexistent target, field written but never read). Dispatched directly when a user wants a flow trac

• implementation-reviewer

  Use this agent after completing all tasks in an implementation plan, or when the user says 'review implementation', 'review changes', 'review and validate', 'execution review', 'audit implementation'. Requires an implementation plan in the current session or a plan file to audit against. Do NOT use

• plan-verifier

  Use this agent to verify an implementation plan before execution. Applies falsifiable error candidates, failure reverse reasoning, design faithfulness checks, and architecture review. Examples: <example> Context: A plan has been written and needs validation before execution. user: "Verify the plan b

• test-runner

  Runs the project's build, test, and lint suites. Detects project type automatically or reads .claude/test-config.yml for overrides. Filters output to errors only with context lines, summary, and return codes. Writes a structured report. Examples: <example> Context: After plan execution completes, ru

Hooks (5)

• SessionStart

  bash ${CLAUDE_PLUGIN_ROOT}/hooks/check-workflow-state.sh

• PreToolUse

  bash ${CLAUDE_PLUGIN_ROOT}/hooks/scan-secrets.sh

• PreToolUse

  python3 ${CLAUDE_PLUGIN_ROOT}/hooks/guard-state-file.py

• PreToolUse

  python3 ${CLAUDE_PLUGIN_ROOT}/hooks/bug-fix-gate.py

• PreToolUse

  python3 ${CLAUDE_PLUGIN_ROOT}/hooks/guard-state-file.py

Similar plugins