본문으로 건너뛰기
cccskills
무료GitHub에 공개

development-harness

Routes to the correct dh skill entry point by intent. Use when unsure which development-harness skill to invoke, starting a development workflow, or invoking /dh directly. Covers capture, groom, plan, execute, single task, quality gates, and milestone routing.

한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.

설치 방법 보기

포함된 파일(2)

  • SKILL.md6.4 KB
  • agents/openai.yaml222 B

SKILL.md(원문)

설치하기 전에 에이전트에게 전달되는 지시 내용을 확인할 수 있습니다.

Development Harness — Plugin Overview and Skill Router

This skill routes to the correct entry point for the development lifecycle. Read it to decide which skill to invoke — not to execute work.

SAM Workflow Pipeline

/dh:add-new-feature  ──>  /dh:implement-feature  ──>  /dh:complete-implementation
   (planning)            (execution loop)         (quality gates)

What This Plugin Provides

The development-harness plugin implements the structured development lifecycle for tracked backlog items. It spans capture through verified closure using a chain of skills backed by the configured backlog backend provider as the source of truth and ~/.dh/projects/{slug}/ as the local state directory. The repository's default configuration currently selects GitHub Issues.

Skills available: /dh:work-backlog-item, /dh:add-new-feature, /dh:implement-feature, /dh:complete-implementation, /dh:gate-push, /dh:work-milestone

Plugin-level source copies exist at plugins/development-harness/skills/ for each skill.


Skill Router — "I want to do X"

flowchart TD
    Start([What do you want to do?]) --> Q1{Intent?}

    Q1 -->|Capture new work —<br>bug, feature idea, observation| Create["/dh:work-backlog-item create<br>Modes: guided intake, quick title, --auto title<br>Writes to ~/.dh/projects/{slug}/backlog/"]

    Q1 -->|Prepare an item for planning —<br>verify claims, map impact, estimate effort| Groom["/dh:work-backlog-item groom {title|section|all}<br>RT-ICA + parallel swarm: fact-checker,<br>impact-analyst, rtica-assessor, classifier, groomer<br>Requires: item exists in backlog"]

    Q1 -->|Plan AND execute a backlog item<br>end-to-end through closure| Work["/dh:work-backlog-item {title|#N|--auto}<br>Handles: auto-groom, RT-ICA gate, SAM planning,<br>GitHub sync, close, resolve<br>STOPS if item already has a plan address"]

    Q1 -->|Plan a feature — produce SAM artifacts<br>without executing| Plan["/dh:add-new-feature {feature description}<br>Phases: discovery → codebase analysis →<br>architecture spec → task decomposition →<br>validation → context manifest<br>Output: feature slug + P{id} task plan"]

    Q1 -->|Execute an existing plan —<br>task plan already produced| Execute["/dh:implement-feature {plan address or slug}<br>Loops ready tasks, dispatches agents,<br>calls complete-implementation when all tasks COMPLETE"]

    Q1 -->|Work a single specific task<br>inside an existing plan| Single["/dh:start-task {plan-address} --task {task-id}<br>Used by implement-feature per-task dispatch —<br>invoke directly to target one task"]

    Q1 -->|Run quality gates after<br>all tasks are COMPLETE| QG["/dh:complete-implementation {plan address|#N}<br>7-task SAM path (with plan): multi-perspective review →<br>code review → verification → integration →<br>doc drift → doc update → context refinement<br>or 5-task proportional path (issue only), which omits<br>multi-perspective review and context refinement"]

    Q1 -->|Work a full milestone<br>in parallel isolated worktrees| Milestone["/dh:work-milestone<br>Wave-based parallel execution — each item<br>gets its own worktree. Use /dh:groom-milestone first."]

Lifecycle — Creation to Verified Closure

flowchart TD
    Capture["/dh:work-backlog-item create<br>Stores the item in the configured backend<br>(default backend: GitHub issue)<br>Status: needs-grooming"] --> Groom
    Groom["/dh:work-backlog-item groom<br>Swarm: impact-analyst, fact-checker,<br>rtica-assessor, classifier, groomer<br>Status: needs-grooming → groomed"] --> Work
    Work["/dh:work-backlog-item work<br>Status: in-progress → discovery gate →<br>groom check → RT-ICA gate → feasibility gate →<br>dh:add-new-feature → plan address on the item"] -->|auto mode| Execute
    Execute["/dh:implement-feature<br>Work loop per ready task: dispatch →<br>runner runs finish → orchestrator settles →<br>orchestrator accepts or reclaims"] --> QG
    QG["/dh:complete-implementation<br>Quality gate plan (7 tasks; 5 without a plan) →<br>status:verified → final commit with Fixes #N, push →<br>backlog resolve"] --> Done(["Status: done"])

    Work -.->|item already has a plan address| Execute
    Work -.->|interactive mode| Stop(["Stops after planning"])
    Close["/dh:work-backlog-item close<br>Dismiss with a reason"] --> Closed(["Status: closed"])
    Resolve["/dh:work-backlog-item resolve<br>Needs status:verified when the item has a plan"] --> Done

Key invariants:

  • When the item already has a plan address, /dh:work-backlog-item runs /dh:implement-feature with that address and stops.
  • When the RT-ICA gate returns BLOCKED, /dh:work-backlog-item sets the item status to blocked and stops.
  • Only the final commit in /dh:complete-implementation carries the Fixes #N trailer; task-level commits during /dh:implement-feature omit it.
  • The SubagentStop hook settles an attempt that the orchestrator did not settle. The hook writes no task status.
  • For an item with a plan, /dh:work-backlog-item resolve requires the status:verified label. The --force flag bypasses this check.

Quick Decision Reference

SituationSkill
Item does not exist yet/dh:work-backlog-item create
Item exists, not yet groomed/dh:work-backlog-item groom {title}
Item is groomed, no plan yet/dh:work-backlog-item {title}
Item has a plan address/dh:implement-feature {plan address or slug}
Plan is executing, one task needs focus/dh:start-task {plan} --task {id}
All tasks complete, run quality gates/dh:complete-implementation {plan address}
Issue number, no plan/dh:complete-implementation #{N} (proportional gates)
Groomed item, skip to planning directly/dh:add-new-feature {description} then /dh:implement-feature
Full milestone in parallel worktrees/dh:groom-milestone then /dh:work-milestone
Dismiss without completing/dh:work-backlog-item close {title}
Mark completed with evidence/dh:work-backlog-item resolve {title}

리뷰

아직 리뷰가 없습니다. 사용해 본 소감을 남겨 주세요.

댓글

로그인 없이 쓸 수 있습니다. 사용 후기, 질문, 작동하지 않은 환경 등을 편하게 남겨 주세요.

아직 댓글이 없습니다. 첫 댓글을 남겨 보세요.

보내면 바로 공개됩니다. 이메일 주소나 전화번호 같은 개인 정보는 쓰지 마세요. URL은 2개까지입니다.

0 / 1,000자

이 양식은 reCAPTCHA로 보호되며 Google 개인정보처리방침 및 서비스 약관이 적용됩니다.

같은 저장소의 스킬

Add automated documentation updater to any Claude skill. Creates a Python sync script that downloads upstream docs, processes markdown for AI consumption, and maintains local cache with configurable refresh. Collects template variables, then delegates implementation through 5-phase workflow. Use when adding auto-updating reference documentation to plugins or skills.

한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.

Jamie-BitFlight/claude_skills672026년 10월 20일 업데이트

SAM-style feature initiation workflow — discovery through codebase analysis, architecture spec, task decomposition, validation, and context manifest. Use when a user asks to add a feature, plan a feature, or convert an idea into an executable SAM plan.

한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.

Jamie-BitFlight/claude_skills672026년 10월 20일 업데이트

Browser automation CLI for AI agents. Use when the user needs to interact with websites, including navigating pages, filling forms, clicking buttons, taking screenshots, extracting data, testing web apps, or automating any browser task. Triggers include requests to "open a website", "fill out a form", "click a button", "take a screenshot", "scrape data from a page", "test this web app", "login to a site", "automate browser actions", or any task requiring programmatic web interaction.

한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.

Jamie-BitFlight/claude_skills672026년 10월 20일 업데이트

Browser automation for AI agents using the agent-browser CLI and Playwright. Use when navigating pages, filling forms, clicking buttons, taking screenshots, extracting data, testing web apps, logging into sites, or automating any browser task. Triggers on "open a website", "fill out a form", "click a button", "scrape data", "test this web app", "automate browser actions", or any programmatic web interaction request.

한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.

Jamie-BitFlight/claude_skills672026년 10월 20일 업데이트

Runs the description-drift experiment — spawns all Claude Code agents simultaneously to collect self-reported capabilities, then compares them against static frontmatter descriptions to reveal how reliable orchestrator routing based on descriptions actually is. Use when measuring description drift across the agent fleet, re-running the capability collection experiment, analyzing a specific agent's self-reported capabilities, or auditing whether frontmatter descriptions accurately reflect agent behavior.

한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.

Jamie-BitFlight/claude_skills672026년 10월 20일 업데이트

Create or adapt Claude Code agent definitions. Use when creating an agent, changing subagent configuration, selecting agent scope, or designing a specialized delegation role.

한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.

Jamie-BitFlight/claude_skills672026년 10월 20일 업데이트

Jamie-BitFlight의 스킬 모두 보기

이 스킬의 문제 신고