{
  "markdown": "![version](https://img.shields.io/badge/version-3.3.0-blue) ![license](https://img.shields.io/badge/license-GPL--3.0-blue) ![Claude Code](https://img.shields.io/badge/Claude%20Code-compatible-orange) ![Cowork](https://img.shields.io/badge/Cowork-compatible-purple)\n\n[Walkthrough](WALKTHROUGH.md) · [Methodology](#methodology) · [Examples](#examples) · [Changelog](CHANGELOG.md) · [Issues](https://github.com/3rd-Brain/AI-Operations-Consultant/issues)\n\n# AI Operations Consultant\n\nA consultant that produces build specs your coding agent can ship from.\n\nFor founders and operators with an automation backlog they haven't shipped. The consultant runs a session and produces a build spec your coding agent can execute against.\n\n## Quick Start\n\n```bash\ngit clone https://github.com/3rd-Brain/AI-Operations-Consultant.git\nmkdir -p ~/.claude/skills\ncp -r AI-Operations-Consultant/skills/ai-ops ~/.claude/skills/\n```\n\nThen invoke `/ai-ops` in any Claude Code or Cowork session. Tell it what you're shipping, where to level up, or what needs documenting. Drop in meeting transcripts, recorded calls, walkthroughs, or any existing docs - the consultant pulls vocabulary and context from them directly, so you don't retype what you've already said. It builds out a living company library and produces specs your coding agents or developers can build from.\n\n## What You Get\n\nEvery session writes to a per-company library you keep and grow over time.\n\n- **Company profile** - what you do, who's on the team, what stack you run\n- **Data architecture** - your records mapped across tools, with sources of truth and fragmentation flagged\n- **Workflow SOPs** - step-by-step playbooks any teammate can run\n- **Role docs** - responsibilities by cadence + KOODAR breakdown per seat\n- **Roadmaps** - maturity-progression plan with priority projects and acceptance signals\n- **Build scopes** - executable specs your coding agent can ship from in one sitting\n\nPlus auto-captured `decisions/` (significant choices, dated) and `glossary.md` (your business's vocabulary).\n\n### Navigator — see your library\n\nThe skill ships with a zero-dependency local web UI (`skills/ai-ops/scripts/navigator/`) that renders the library as a wiki + dashboard: an org chart built from your role docs, color-coded step maps for every workflow SOP, live dashboard facts derived from the files (counts, completeness, top pains, maturity signal), and in-place editing of any page. Ask the consultant to \"open the navigator,\" or run it yourself:\n\n```bash\nROOT=\"<your-company-library>\" node ~/.claude/skills/ai-ops/scripts/navigator/server.mjs --port 4173\n```\n\nLocalhost only, markdown writes only, no separate data store — it reads and writes the same library files.\n\n## How It Works\n\nYou have a conversation. The consultant dispatches specialists when work benefits from a fresh context window. The deliverables come back as files in a per-company library you own.\n\n```\n   user\n    │\n    ▼\n┌─────────────────────────┐\n│  /ai-ops                │   conversational orchestrator\n│(one skill, one session) │\n└──────────┬──────────────┘\n           │  dispatches specialists in subagents\n           ▼\n┌──────────────────────────────────────────────────────────┐\n│  Workflow chain   Company chain   Role   Roadmap   Build  │\n│  (SOP / tools /   (profile /      Scope                   │\n│   automation)      stack research /                       │\n│                    data arch)                             │\n└──────────┬───────────────────────────────────────────────┘\n           │  writes files to the per-company library\n           ▼\n       <company>/\n         profile.md · data-architecture.md · glossary.md\n         workflows/ · roles/ · tools/\n         decisions/ · roadmaps/ · scopes/\n```\n\nThe library is plain markdown so you can read and edit any file directly.\n\n## Methodology\n\nFive frameworks the consultant applies. Source: the Digital Operations Playbook (bundled in the skill at `references/playbook/`).\n\n- **5 Levels of Digital Operations** - where you sit on the maturity scale (Information Silos → Connectable Cloud → Unified Data Layer → Automated Workflows → AI Automation). Level 3 is the flywheel threshold.\n- **Three Cs** - Consistency → Clarity → Capacity. The chain reaction you unlock as you level up. Each one is a prerequisite for the next.\n- **5 Pillars** - Talent Strategy, Workflow Optimization, Digital Architecture, Knowledge Management, AI Automation. What you actually build to move the levels.\n- **People → Process → Tools** - the implementation sequence within every pillar. Skip it and you ship shelfware.\n- **KOODAR** - the AI capabilities vocabulary (Know / Observe / Orient / Decide / Act / Review). The language for designing and documenting AI tools and agents.\n\n## What to ask for and what the consultant produces.\n\n| What you say                                           | What you get                                                                                                                                         |\n| ------------------------------------------------------ | ---------------------------------------------------------------------------------------------------------------------------------------------------- |\n| \"Map how our business runs\"                            | The full picture: a company profile, cross-tool data architecture, and entries for every tool you operate on                                         |\n| \"Walk through our quoting workflow\"                    | A step-by-step SOP a teammate can follow, plus research on every tool in the workflow and the automation opportunities buried in it                  |\n| \"Document the head of sales role\"                      | A role doc with responsibilities by cadence, objectives and metrics, and a KOODAR capability breakdown                                               |\n| \"Give me a roadmap to get our marketing team AI ready\" | A sequenced plan: priority projects, acceptance signals, scope-specific pitfalls, and the library work that needs to happen first                    |\n| \"Scope the carrier-quote extractor\"                    | A build spec your coding agent can ship from: stack picks, input/output contracts, tool docs and APIs, examples, and test-driven acceptance criteria |\n\nAnything mentioned but not yet in the library - a tool, a role, a workflow - gets a stub file created automatically. The library grows with every session.\n\n## Compatibility\n\nThe skill ships as a folder (`skills/ai-ops/`) and a zipped bundle (`dist/cowork/ai-ops.skill`). All five supported runtimes have native skill systems that accept the same SKILL.md + subfolder structure.\n\n| Agent | Skills directory | Install |\n|---|---|---|\n| **Claude Code** | `~/.claude/skills/` | `cp -r skills/ai-ops ~/.claude/skills/` |\n| **Cowork** | (managed in-app) | Import `dist/cowork/ai-ops.skill` via **Skills → Create skill → Import** |\n| **OpenClaw** | `~/.openclaw/skills/` | `cp -r skills/ai-ops ~/.openclaw/skills/` or `unzip dist/cowork/ai-ops.skill -d ~/.openclaw/skills/` |\n| **Codex CLI** (OpenAI) | `~/.agents/skills/` | `cp -r skills/ai-ops ~/.agents/skills/` or `unzip dist/cowork/ai-ops.skill -d ~/.agents/skills/` |\n| **Hermes** (NousResearch) | `~/.hermes/skills/` | `hermes skills install https://github.com/3rd-Brain/AI-Operations-Consultant/raw/main/dist/cowork/ai-ops.skill` |\n\nAll five read SKILL.md YAML frontmatter (`name`, `description`). The `license` key is Claude Code / Cowork only - silently ignored by the others.\n\nFor any other Claude or OpenAI-compatible runtime, point it at `skills/ai-ops/SKILL.md` as the system prompt.\n\n## Examples\n\n### Document an inquiry-to-quote workflow\n\n> \"Walk me through how new event inquiries become signed quotes. We use Good Shuffle, HoneyBook, and Gmail.\"\n\nThe consultant dispatches the workflow chain. Output:\n- `workflows/inquiry-to-quote.md` - full SOP with steps, tools per step, escalation paths, and an Automation / Agent Potential section\n- `.ai-ops/research/inquiry-to-quote-tools-2026-05-21.md` - API capabilities, native integrations, integration walls\n- Stub files created for any tool / role / record mentioned but not yet in the library\n\n### Plan a level-up roadmap for the proposal workflow\n\n> \"I want a roadmap for getting our proposal workflow team-shared so any teammate can produce a consistent output. We're at Level 2 today.\"\n\nThe consultant dispatches the Roadmap specialist. Output:\n- `roadmaps/proposal-workflow-team-shared-<date>.md` - operator goal, three Priority projects, two Upcoming projects, acceptance signals, scope-specific pitfalls, next-action library work\n\n### Scope a build for the highest-priority automation\n\nAfter the workflow chain runs, ask:\n\n> \"Scope the carrier-quote PDF extractor. The one from our discussion in /Meetings from yesterday.\"\n\nThe consultant walks stack selection (Python + Pydantic? Make scenario? Existing Airtable? Each slot one question at a time), then dispatches Build Scope. Output:\n- `scopes/carrier-quote-pdf-extractor-<date>.md` - what to build, the chosen stack, input/output contracts, tools and APIs, records read/written, implementation notes, test-driven acceptance criteria\n\nDrop the scope into Claude Code, Cowork, or hand to a developer.\n\n## Configuration\n\nThe consultant works out of the box. Per-session things you can tune in conversation:\n\n- **Library location** - where your per-company files live (default: a folder you specify when the session opens)\n- **Existing research** - if you've already gathered tool API docs or run prior research, drop the files in `<library>/.ai-ops/research/` and the consultant picks them up\n\nNo config file required. The library structure carries everything.\n\n## Community & Support\n\n- **Issues:** [github.com/3rd-Brain/AI-Operations-Consultant/issues](https://github.com/3rd-Brain/AI-Operations-Consultant/issues)\n- **Methodology source:** [digitaloperationsinstitute.com](https://digitaloperationsinstitute.com)\n- **Behind the methodology:** [3rdbrain.co](https://3rdbrain.co) - the consulting firm that built it across 50+ engagements\n\n## License\n\nGPL-3.0 - see [LICENSE](LICENSE).\n",
  "bytes": 10131,
  "sha": "04d8040217c4fde275643e5e47f2e4f1f342322e637995b00cd74a57848783ab",
  "repo_slug": "3rd-brain/ai-operations-consultant",
  "fonte": "repo",
  "truncated": false,
  "api": "https://agentalog.com/api/listings/plg_3rd_brain_ai_operations_consultant_ai_op_d30afcb6/readme"
}