[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"project-93274":3},{"id":4,"name":5,"fullName":6,"owner":7,"repo":5,"description":8,"homepage":9,"htmlUrl":9,"language":10,"languages":9,"totalLinesOfCode":9,"stars":11,"forks":12,"watchers":13,"openIssues":14,"contributorsCount":14,"subscribersCount":14,"size":14,"stars1d":14,"stars7d":15,"stars30d":16,"stars90d":14,"forks30d":14,"starsTrendScore":14,"compositeScore":17,"rankGlobal":9,"rankLanguage":9,"license":18,"archived":19,"fork":19,"defaultBranch":20,"hasWiki":21,"hasPages":19,"topics":22,"createdAt":9,"pushedAt":9,"updatedAt":31,"readmeContent":32,"aiSummary":33,"trendingCount":14,"starSnapshotCount":14,"syncStatus":34,"lastSyncTime":35,"discoverSource":36},93274,"sample-specship","aws-samples\u002Fsample-specship","aws-samples","Spec-driven autonomous engineering workflow for AI coding agents: recon → plan → build → validate → ship — with TDD, adversarial validation, and anti-slop quality gates. Packaged as a Kiro Power.",null,"Shell",160,1,141,0,3,20,44.4,"MIT License",false,"main",true,[23,24,25,26,27,28,29,30],"ai-agents","autonomous-agents","code-generation","developer-tools","kiro","llm","spec-driven-development","tdd","2026-07-23 04:02:09","# SpecShip\n\nA [Kiro Power](https:\u002F\u002Fkiro.dev) autonomous engineering workflow that orchestrates AI agents through a complete **recon → plan → build → validate → ship** pipeline — with brownfield reverse engineering, TDD, parallel execution, adversarial validation, and spec-driven quality gates.\n\n## Architecture\n\n![SpecShip architecture](docs\u002Farchitecture-diagram.png)\n\nEditable source: [`docs\u002Farchitecture-diagram.svg`](docs\u002Farchitecture-diagram.svg)\n\n## The 4 Phases (+ Brownfield Recon)\n\n| Phase | What happens | Output |\n|-------|-------------|--------|\n| **RECON** *(brownfield only)* | Reverse engineer existing code before planning: repo map, technology stack, components, APIs, data models, business flows, baseline tests, preserved behavior, and change impact. | `.specship\u002F...\u002Fartifacts\u002Freverse-engineering\u002F` |\n| **PLAN** | Brainstorm → market research (real web search, minimum 5 queries) → sprint contract → pre-generate test cases → implementation plan. Each step feeds the next. | `requirements.md` + `design.md` + `tasks.md` + artifacts |\n| **BUILD** | Execute tasks.md milestone-by-milestone. Parallel batches for independent tasks. Every behavior follows Test-Driven Development (write failing test → watch it fail → implement → pass → commit). Hard gate blocks each milestone until typecheck + tests + build pass. | Tested, committed code per milestone |\n| **VALIDATE** | Up to 7 adversarial validators run in parallel as independent subagents, each delegating to its gstack specialist skill (code→review, security→cso, browser→qa-only). Code, Security, Integration, Browser, Design, Alignment (always) + Load (if performance NFR exists). Each produces a typed verdict with evidence. Aggregate decides: merge \u002F recover \u002F escalate. | Typed verdicts → merge or recover |\n| **SHIP** | PR + changelog + archive. Ship report with timing, bugs caught, milestone breakdown. If no git remote, prepares locally with push instructions. | PR ready to merge |\n\n## Why SpecShip?\n\nAI agents build fast but shallow — working skeleton in 7 minutes, 60% of features missing, no tests, no error handling. SpecShip fixes this:\n\n| Principle | What it means |\n|-----------|--------------|\n| **Test-Driven Development** | No production code without a failing test first. The build gate blocks every milestone until typecheck + tests + build all pass. Catches bugs within minutes, not hours. |\n| **Parallel execution** | Independent tasks run simultaneously. All validators fire in parallel as independent subagents. An hour of sequential work finishes in minutes. |\n| **Adversarial validation** | The agent that built the code CANNOT judge it. Independent validators produce typed verdicts with evidence. A separate skill with its own methodology judges the code. |\n| **Self-healing recovery** | When a validator finds a bug, a fresh agent fixes it surgically — one fix per issue, regression test first, max 3 cycles. Parallel fixes when touching different files. |\n| **Contract-first** | Every build starts with acceptance criteria + failure modes + design spec — defined before any code and stored in `.specship\u002Fspecs\u002F\u003Cid>\u002Frequirements.md` + `design.md`. Validators judge against the contract, not vibes. |\n| **Brownfield-safe** | Existing code gets reverse-engineered before planning. The contract includes preserved behavior, local conventions, baseline tests, files likely to change, and files to avoid. |\n| **Market research** | Conducts real web searches (minimum 5 queries) to study existing products in the same category. Sets the quality bar from what exists — not from training data or tutorials. Research is saved to `.specship\u002Fspecs\u002F\u003Cid>\u002Fartifacts\u002Fmarket-research.md`. |\n\n## Install\n\n```bash\n.\u002Finstall.sh\n```\n\nThat's it. The installer automatically:\n- Detects and installs missing dependencies (superpowers + gstack)\n- Copies SpecShip steering files into `~\u002F.kiro\u002Fsteering\u002F`\n- Detects if Playwright MCP is configured — if not, sets it up (including Node.js guidance if missing)\n\nOne command, everything set up.\n\n> By default installs globally (`~\u002F.kiro\u002Fsteering\u002F`). To scope to a single project: `.\u002Finstall.sh .\u002F.kiro\u002Fsteering`\n\n## Use\n\nJust talk to Kiro — the `auto` skills match your intent:\n\n```\nUsing SpecShip, build me a Kanban board\nreverse engineer this repo before adding billing\nstart building\nvalidate\nship it\nwhere was I?            ← resumes an interrupted mission\n```\n\nWant to force a specific skill? Type its slash command (`\u002Fspecship-plan`, `\u002Fspecship-validate`) or `#`-reference it (`#specship-plan`).\n\n## What's in it\n\n```\nkiro-power-specship\u002F\n├── power.json                          # Power manifest\n├── POWER.md                            # Full documentation (shown on Try power)\n├── PREREQUISITES.md                    # Required companions (superpowers + gstack) + Playwright MCP\n├── SECURITY.md                         # Supply-chain notes + how to report an issue\n├── install.sh                          # Installer (auto-installs dependencies + Playwright MCP)\n├── uninstall.sh                        # Removes exactly what install.sh added\n├── build-cli-skills.sh                 # Generates CLI skills from steering files\n├── settings\u002Fmcp.json                   # Playwright + Chrome DevTools MCP config (--isolated)\n├── settings\u002Fmcp\u002F                        # Pinned MCP servers + integrity lockfile (npm ci)\n├── specship-verify.sh                   # Integrity verifier (detects tampered steering\u002Fskills)\n├── hooks\u002F                              # Kiro agent hooks (opt-in, disabled by default)\n├── templates\u002Fspecship-gitignore        # .gitignore template for .specship\u002F folder\n├── docs\u002F                               # Architecture diagrams (PNG + SVG + Mermaid source)\n└── steering\u002F\n    ├── specship-workflow.md            # [always] pipeline + routing + enforcement rules\n    ├── specship-guardrails.md          # [always] 19 non-negotiable rules\n    ├── specship-prerequisites.md       # [always] companion delegation map\n    ├── specship-reverse-engineer.md    # [auto] brownfield repo reconnaissance\n    ├── specship-plan.md                # [auto] plan pipeline orchestrator\n    ├── specship-contract.md            # [auto] sprint contract generation\n    ├── specship-testgen.md             # [auto] pre-write tests before code\n    ├── specship-build.md               # [auto] milestone loop + parallel batches\n    ├── specship-validate.md            # [auto] validation orchestrator (parallel subagents)\n    ├── specship-recover.md             # [auto] surgical fixes, regression-test-first\n    ├── specship-ship.md                # [auto] PR + changelog + archive\n    ├── specship-resume.md              # [auto] resume interrupted missions\n    ├── specship-validate-*.md (×8)     # [manual] individual validator skills\n    └── shared\u002F (×12)                   # [manual] reference docs pulled on demand\n```\n\n## Modes\n\n| Mode | Trigger | Behavior |\n|------|---------|----------|\n| **Guided** (default) | `Using SpecShip, build me X` | Runs the pipeline, pauses for plan approval before building |\n| **Autonomous** | `SpecShip auto: build me X` or `build me X, fully autonomous` | Runs the entire pipeline end-to-end without stopping. You come back to a PR. |\n\n## Customize\n\nTreat SpecShip as a baseline and build your own workflow on top of it. To update, re-run `.\u002Finstall.sh` (it backs up any files you've edited before overwriting). To remove everything, run `.\u002Funinstall.sh`.\n\nExperimental and unofficial. SpecShip has not undergone external security review and is provided as-is, with no warranty — use at your own risk. All generated code is sample\u002Freference implementation requiring AppSec review before production use. Licensed under the [MIT License](LICENSE).\n","SpecShip 是一个面向 AI 编程代理的规范驱动型自主工程工作流，实现从逆向工程（RECON）到交付（SHIP）的完整闭环：支持存量系统反向建模、契约先行的计划生成、测试驱动的并行构建、多角色对抗式验证（代码\u002F安全\u002F浏览器等维度），以及自动 PR 交付。技术上依托 Kiro Power 框架，强调 TDD 强约束、子代理并行执行与证据可追溯的验证 verdict。适用于需要高可信度 AI 辅助开发的中大型软件工程场景，尤其适合遗留系统演进、合规敏感型项目及需审计追踪的自动化研发流程。",2,"2026-07-15 02:30:06","CREATED_QUERY"]