{
  "markdown": "<p align=\"center\">\n  <picture>\n    <source media=\"(prefers-color-scheme: dark)\" srcset=\"assets/yamadaaa.png\">\n    <img src=\"assets/yamadaaa.png\" width=\"350\" alt=\"Iterative Dev Workflow\">\n  </picture>\n</p>\n\n<h1 align=\"center\">Iterative Dev Workflow</h1>\n\n<p align=\"center\">\n  <em>Structure your AI agent's development process.</em>\n</p>\n\n<p align=\"center\">\n  <sub>METHODOLOGY &amp; PROMPT DESIGN</sub>\n</p>\n\n<p align=\"center\">\n  <a href=\"https://github.com/haniladjamba\">\n    <img src=\"https://github.com/haniladjamba.png\" width=\"64\" height=\"64\" style=\"border-radius:50%\" alt=\"Hani Ladjamba\"/>\n  </a>\n  &nbsp;&nbsp;&nbsp;\n  <a href=\"https://github.com/tiwowtimothy\">\n    <img src=\"https://github.com/tiwowtimothy.png\" width=\"64\" height=\"64\" style=\"border-radius:50%\" alt=\"Timothy Tiwow\"/>\n  </a>\n</p>\n<p align=\"center\">\n  <a href=\"https://github.com/haniladjamba\"><b>Hani Ladjamba</b></a>\n  &nbsp;&nbsp;&nbsp;&nbsp;\n  <a href=\"https://github.com/tiwowtimothy\"><b>Timothy Tiwow</b></a>\n</p>\n\n---\n\nStop shipping features that break production. Stop rewriting the same architecture three times. Stop wondering why your AI agent keeps over-engineering simple tasks.\n\nIterative Dev Workflow gives your AI agent a structured development process — from understanding the task to reflecting on what went wrong.\n\n## Before / after\n\nYou give your AI agent a vague task like \"build user authentication.\" Without a structure, it writes code, asks no questions, and delivers something that doesn't match what you needed.\n\nWith Iterative Dev Workflow:\n\n```\n/kickoff\n  → Agent asks: what platform? what stack? what's the project name?\n  → You answer\n  → Agent confirms understanding, sets up .workflow/<slug>/00-context.md\n\n/phase-1\n  → Agent explores codebase, identifies gaps, asks the right questions\n\n/phase-2-step-1\n  → Agent proposes 5 different approaches (Architecture, UI/UX, Data Model, etc.)\n\n/phase-2-step-2\n  → Agent analyzes trade-offs and sensitivity points for each approach\n\n/phase-2-step-3\n  → You define Quality Attributes (e.g. Performance, Maintainability)\n  → Agent scores each design against your criteria\n\n/phase-2-step-4\n  → You choose the best design\n  → Agent creates System Context Diagram (Mermaid.js)\n\n/phase-2-step-5\n  → Agent creates User Journey Diagram (Mermaid.js)\n\n/phase-3\n  → Agent writes implementation plan — avoiding overengineering\n\n/phase-4\n  → Agent reflects: what went well, what could be better\n```\n\nYou stay in control. The agent never skips ahead.\n\n## Install\n\n### Claude Code\n\n```\n/plugin marketplace add arsxxi/iterative-dev-workflow\n/plugin install iterative-dev-workflow@arsxxi-iterative-dev-workflow\n```\n\n### Codex\n\n```bash\ncodex plugin marketplace add arsxxi/iterative-dev-workflow\ncodex plugin add iterative-dev-workflow@arsxxi-iterative-dev-workflow\n```\n\n### OpenCode\n\n```bash\nnpm install -g @arsxxi/iterative-dev-workflow\n```\n\nOn install, a `postinstall` script copies the commands into `~/.config/opencode/commands/`\nautomatically. Restart OpenCode and type `/` to see them.\n\nIf commands still don't show up (some package managers or environments skip lifecycle scripts,\nor your OpenCode version doesn't pick them up automatically), run the installer manually:\n\n```bash\nnpx --package=@arsxxi/iterative-dev-workflow iterative-dev-workflow-install\n```\n\nOr, as a guaranteed last resort, copy the `commands/` folder from this repo directly into\n`~/.config/opencode/commands/` (global) or `.opencode/commands/` inside your project yourself -\nthese are plain markdown files, no build step required.\n\nThen add to your `opencode.json` (this enables the AGENTS.md system-prompt injection feature,\nseparate from command registration):\n\n```json\n{ \"plugin\": [\"@arsxxi/iterative-dev-workflow\"] }\n```\n\n### Kilo Code\n\nKilo Code is a VS Code extension, so there is no plugin marketplace command — the commands are\ninstalled as files. Easiest path, available in every project:\n\n```bash\nnpm install -g @arsxxi/iterative-dev-workflow\nnpx --package=@arsxxi/iterative-dev-workflow iterative-dev-workflow-install-kilo\n```\n\nThis copies the 10 commands into `~/.config/kilo/commands/` and the methodology into\n`~/.config/kilo/rules/`. Reload the Kilo Code extension, then type `/` to see them.\n\nTo load the methodology into the system prompt, add the rules glob to\n`~/.config/kilo/kilo.jsonc`:\n\n```json\n{ \"instructions\": [\"~/.config/kilo/rules/*.md\"] }\n```\n\nFor a single project instead of globally, copy `.kilo/commands/` and `.kilo/rules/` from this\nrepo into your project root and point `kilo.jsonc` at the rules:\n\n```json\n{ \"instructions\": [\".kilo/rules/*.md\"] }\n```\n\nOlder Kilo Code builds read `.kilocode/workflows/` and `.kilocode/rules/` instead. Both paths are\nshipped in this repo, so either version works — newer builds migrate the legacy path on startup.\n\n### Antigravity CLI\n\n```bash\nagy plugin install https://github.com/arsxxi/iterative-dev-workflow\n```\n\n## Commands\n\n| # | Command | Description |\n|---|---------|-------------|\n| 0 | `/kickoff` | Start a new project — asks what to build, platform/stack, project name |\n| 1 | `/phase-1` | Analyze: understand task, explore codebase, identify gaps |\n| 2.1 | `/phase-2-step-1` | Solution Proposal: create minimum 5 designs |\n| 2.2 | `/phase-2-step-2` | ATAM: assess trade-offs and sensitivity points |\n| 2.3 | `/phase-2-step-3` | Quality Attribute: weighted scoring assessment |\n| 2.4 | `/phase-2-step-4` | High-Fidelity Design: System Context Diagram (Mermaid.js) |\n| 2.5 | `/phase-2-step-5` | User Journey: User Journey Diagram (Mermaid.js) |\n| 3 | `/phase-3` | Implementation Plan: write comprehensive plan |\n| 4 | `/phase-4` | Postmortem: reflect and improve |\n| — | `/session-transcript` | Record verbatim conversation to project root as `aichat-<slug>.md` |\n\n## Output Location\n\nAll phase outputs are written to `.workflow/<slug>/`:\n\n```\n.workflow/<slug>/\n├── 00-context.md      # platform, description, existing services\n├── 01-analyze.md     # Phase 1\n├── 02-propose.md     # Phase 2 Step 1\n├── 02-atam.md        # Phase 2 Step 2\n├── 02-qa.md          # Phase 2 Step 3\n├── 02-hifi.md        # Phase 2 Step 4 (System Context Diagram)\n├── 02-journey.md     # Phase 2 Step 5 (User Journey Diagram)\n├── 03-implement.md    # Phase 3\n└── 04-postmortem.md   # Phase 4\n```\n\nSession transcripts are saved directly to the **project root** as `aichat-<slug>.md` (or `aichat.md` if no project name was provided).\n\n## Hard Constraints\n\n- **AVOID overengineering.** PREFER simple, low-complexity implementations.\n- **AVOID jargon.** Use plain language that states actual intent.\n- **Never skip phases.** Design must be approved before implementation begins.\n- **Iterative, not waterfall.** If a problem surfaces in a later phase, circle back to fix it there.\n\n## Workflow Flow\n\n```\n/kickoff\n  └── /phase-1\n        └── /phase-2-step-1\n              └── /phase-2-step-2\n                    └── /phase-2-step-3  (may loop back to /phase-2-step-1)\n                          └── /phase-2-step-4\n                                └── /phase-2-step-5\n                                      └── /phase-3\n                                            └── /phase-4\n```\n\n## Development\n\nThe source of truth is `commands/*.md` (10 files) and `skills/workflow-methodology/SKILL.md`. After editing these, run:\n\n```bash\nbash scripts/sync-platforms.sh\n```\n\nThis syncs to:\n- `.opencode/commands/` — OpenCode command definitions\n- `.agents/skills/` — Antigravity/Codex skill definitions\n- `.kilo/commands/` + `.kilo/rules/` — Kilo Code commands and methodology rule\n- `.kilocode/workflows/` + `.kilocode/rules/` — same payload on Kilo Code's legacy paths\n- `AGENTS.md` — cross-platform instruction file\n\nThe Kilo Code targets are the only ones that aren't a verbatim copy: Kilo has no `argument-hint`\nfrontmatter key and doesn't substitute `$ARGUMENTS`, so `scripts/build-kilo.mjs` rewrites the\nfrontmatter and prepends a short note explaining where the argument comes from. The command body\nitself is copied unchanged.\n\n## FAQ\n\n**Does every project need all phases?**\n\nNo. The workflow is modular. Use only what the task needs — but don't skip backward. Each phase has a clear prerequisite chain.\n\n**What if the design doesn't support implementation?**\n\nStop. Tell the user clearly. Circle back to Phase 2. Do not silently work around a design flaw in the plan.\n\n**How do I choose a project name?**\n\nUse a short, lowercase identifier with hyphens (e.g. `user-auth`, `article-quality-widget`, `payment-flow`). It's used as the project folder name under `.workflow/<slug>/`, so it stays filesystem-safe.\n\n**What's the difference between Phase 2 Step 4 and Step 5?**\n\nStep 4 creates a System Context Diagram — shows how the solution fits within the whole app. Step 5 creates a User Journey Diagram — shows how the user interacts with the system.\n\n## License\n\n[MIT](LICENSE)\n",
  "bytes": 8824,
  "sha": "1377028bfd5d1298fb3f2b3f3ab38cc99dd0db89c8283be0295cbd3f4678fbdd",
  "repo_slug": "arsxxi/iterative-dev-workflow",
  "fonte": "repo",
  "truncated": false,
  "api": "https://agentalog.com/api/listings/plg_arsxxi_iterative_dev_workflow_0962f539/readme"
}