Add Codex integration and installer support

README.md

@@ -44,7 +44,7 @@ Each agent file contains:
 
 Browse the agents below and copy/adapt the ones you need!
 
-### Option 3: Use with Other Tools (Cursor, Aider, Windsurf, Gemini CLI, OpenCode)
+### Option 3: Use with Other Tools (Codex, Cursor, Aider, Windsurf, Gemini CLI, OpenCode)
 
 ```bash
 # Step 1 -- generate integration files for all supported tools
@@ -54,6 +54,7 @@ Browse the agents below and copy/adapt the ones you need!
 ./scripts/install.sh
 
 # Or target a specific tool directly
+./scripts/install.sh --tool codex
 ./scripts/install.sh --tool cursor
 ./scripts/install.sh --tool aider
 ./scripts/install.sh --tool windsurf
@@ -324,7 +325,7 @@ Each agent is designed with:
 
 ## 📊 Stats
 
-- 🎭 **61 Specialized Agents** across 9 divisions
+- 🎭 **68 Specialized Agents** across 9 divisions
 - 📝 **10,000+ lines** of personality, process, and code examples
 - ⏱️ **Months of iteration** from real-world usage
 - 🌟 **Battle-tested** in production environments
@@ -339,6 +340,7 @@ The Agency works natively with Claude Code, and ships conversion + install scrip
 ### Supported Tools
 
 - **[Claude Code](https://claude.ai/code)** — native `.md` agents, no conversion needed → `~/.claude/agents/`
+- **Codex** — one meta-skill with internal sub-skills + NEXUS references → `~/.codex/skills/agency-agents/`
 - **[Antigravity](https://github.com/google-gemini/antigravity)** — `SKILL.md` per agent → `~/.gemini/antigravity/skills/`
 - **[Gemini CLI](https://github.com/google-gemini/gemini-cli)** — extension + `SKILL.md` files → `~/.gemini/extensions/agency-agents/`
 - **[OpenCode](https://opencode.ai)** — `.md` agent files → `.opencode/agent/`
@@ -369,20 +371,22 @@ The installer scans your system for installed tools, shows a checkbox UI, and le
 
   System scan: [*] = detected on this machine
 
-  [x]  1)  [*]  Claude Code     (claude.ai/code)
-  [x]  2)  [*]  Antigravity     (~/.gemini/antigravity)
-  [ ]  3)  [ ]  Gemini CLI      (gemini extension)
-  [ ]  4)  [ ]  OpenCode        (opencode.ai)
-  [x]  5)  [*]  Cursor          (.cursor/rules)
-  [ ]  6)  [ ]  Aider           (CONVENTIONS.md)
-  [ ]  7)  [ ]  Windsurf        (.windsurfrules)
+  [x]  1)  [*]  Codex           (~/.codex/skills)
+  [x]  2)  [*]  Claude Code     (claude.ai/code)
+  [x]  3)  [*]  Antigravity     (~/.gemini/antigravity)
+  [ ]  4)  [ ]  Gemini CLI      (gemini extension)
+  [ ]  5)  [ ]  OpenCode        (opencode.ai)
+  [x]  6)  [*]  Cursor          (.cursor/rules)
+  [ ]  7)  [ ]  Aider           (CONVENTIONS.md)
+  [ ]  8)  [ ]  Windsurf        (.windsurfrules)
 
-  [1-7] toggle   [a] all   [n] none   [d] detected
+  [1-8] toggle   [a] all   [n] none   [d] detected
   [Enter] install   [q] quit
 ```
 
 **Or install a specific tool directly:**
 ```bash
+./scripts/install.sh --tool codex
 ./scripts/install.sh --tool cursor
 ./scripts/install.sh --tool opencode
 ./scripts/install.sh --tool antigravity
@@ -397,6 +401,24 @@ The installer scans your system for installed tools, shows a checkbox UI, and le
 
 ### Tool-Specific Instructions
 
+<details>
+<summary><strong>Codex</strong></summary>
+
+The Codex integration packages the full Agency catalog as one public meta-skill
+with internal role sub-skills and NEXUS references.
+
+```bash
+./scripts/install.sh --tool codex
+```
+
+Then invoke it in Codex:
+```
+Use $agency-agents to pick the right Agency specialist for this task.
+```
+
+See [integrations/codex/README.md](integrations/codex/README.md) for details.
+</details>
+
 <details>
 <summary><strong>Claude Code</strong></summary>
 
@@ -434,7 +456,7 @@ See [integrations/antigravity/README.md](integrations/antigravity/README.md) for
 <details>
 <summary><strong>Gemini CLI</strong></summary>
 
-Installs as a Gemini CLI extension with 61 skills + a manifest.
+Installs as a Gemini CLI extension with the full Agency roster + a manifest.
 
 ```bash
 ./scripts/install.sh --tool gemini-cli
@@ -538,7 +560,7 @@ When you add new agents or edit existing ones, regenerate all integration files:
 
 - [ ] Interactive agent selector web tool
 - [x] Multi-agent workflow examples -- see [examples/](examples/)
-- [x] Multi-tool integration scripts (Claude Code, Antigravity, Gemini CLI, OpenCode, Cursor, Aider, Windsurf)
+- [x] Multi-tool integration scripts (Codex, Claude Code, Antigravity, Gemini CLI, OpenCode, Cursor, Aider, Windsurf)
 - [ ] Video tutorials on agent design
 - [ ] Community agent marketplace
 - [ ] Agent "personality quiz" for project matching

integrations/README.md

@@ -1,10 +1,11 @@
 # 🔌 Integrations
 
-This directory contains The Agency's 61 AI agents converted into formats
+This directory contains The Agency's 68 AI agents converted into formats
 compatible with popular agentic coding tools.
 
 ## Supported Tools
 
+- **[Codex](#codex)** — one meta-skill with internal sub-skills in `codex/`
 - **[Claude Code](#claude-code)** — `.md` agents, use the repo directly
 - **[Antigravity](#antigravity)** — `SKILL.md` per agent in `antigravity/`
 - **[Gemini CLI](#gemini-cli)** — extension + `SKILL.md` files in `gemini-cli/`
@@ -21,6 +22,7 @@ compatible with popular agentic coding tools.
 
 # Install for a specific tool
 ./scripts/install.sh --tool antigravity
+./scripts/install.sh --tool codex
 ./scripts/install.sh --tool gemini-cli
 ./scripts/install.sh --tool cursor
 ./scripts/install.sh --tool aider
@@ -38,6 +40,21 @@ If you add or modify agents, regenerate all integration files:
 
 ---
 
+## Codex
+
+The Agency is packaged for Codex as a single public meta-skill named
+`agency-agents`. The skill routes into internal specialist sub-skills under
+`codex/agency-agents/sub/` and includes NEXUS references under
+`codex/agency-agents/references/`.
+
+```bash
+./scripts/install.sh --tool codex
+```
+
+See [codex/README.md](codex/README.md) for details.
+
+---
+
 ## Claude Code
 
 The Agency was originally designed for Claude Code. Agents work natively

integrations/aider/README.md

@@ -1,6 +1,6 @@
 # Aider Integration
 
-All 61 Agency agents are consolidated into a single `CONVENTIONS.md` file.
+All Agency agents are consolidated into a single `CONVENTIONS.md` file.
 Aider reads this file automatically when it's present in your project root.
 
 ## Install

integrations/antigravity/README.md

@@ -1,6 +1,6 @@
 # Antigravity Integration
 
-Installs all 61 Agency agents as Antigravity skills. Each agent is prefixed
+Installs the full Agency roster as Antigravity skills. Each agent is prefixed
 with `agency-` to avoid conflicts with existing skills.
 
 ## Install

integrations/claude-code/README.md

@@ -28,4 +28,4 @@ Use the Reality Checker agent to verify this feature is production-ready.
 ## Agent Directory
 
 Agents are organized into divisions. See the [main README](../../README.md) for
-the full roster of 61 specialists.
+the full roster of specialists.

integrations/codex/README.md

@@ -0,0 +1,52 @@
+# Codex Integration
+
+The Agency is packaged for Codex as one public meta-skill named
+`agency-agents`. The public skill stays small, while all specialists live as
+internal sub-skills under `sub/`, and the NEXUS playbooks/runbooks live under
+`references/`.
+
+## Install
+
+```bash
+./scripts/convert.sh --tool codex
+./scripts/install.sh --tool codex
+```
+
+This copies `integrations/codex/agency-agents/` to
+`~/.codex/skills/agency-agents/`.
+
+## Activate The Skill
+
+In Codex, invoke the skill directly:
+
+```
+Use $agency-agents to choose the right Agency specialist for this repo.
+```
+
+```
+Use $agency-agents and run a NEXUS-style workflow for this feature.
+```
+
+## Skill Layout
+
+```text
+~/.codex/skills/agency-agents/
+  SKILL.md
+  agents/openai.yaml
+  sub/<agent-slug>/subskill.md
+  references/roster.md
+  references/strategy/...
+  references/examples/...
+```
+
+## Why One Skill Instead Of Dozens Of Public Skills
+
+Codex works best when the public skill list stays compact. The Agency therefore
+ships as one public skill that routes into internal role files on demand,
+instead of exposing dozens of separate public skills.
+
+## Regenerate
+
+```bash
+./scripts/convert.sh --tool codex
+```

integrations/codex/agency-agents/SKILL.md

@@ -0,0 +1,56 @@
+---
+name: agency-agents
+description: Meta-skill for The Agency on Codex. Use when the user wants a specialized Agency role, multi-role coordination, or the NEXUS orchestration workflow inside Codex.
+---
+
+# The Agency for Codex
+
+`agency-agents` packages the full Agency roster as one public Codex skill with
+internal role sub-skills under `sub/` and NEXUS references under `references/`.
+
+## When To Use
+
+- The user asks for The Agency, NEXUS, or a named Agency specialist.
+- The work benefits from a temporary specialist persona instead of a generic assistant.
+- The task needs multi-role coordination across product, design, engineering, QA, launch, or operations.
+
+## Operating Model
+
+1. Decide whether the task needs one specialist, a small squad, or a NEXUS flow.
+2. Read `references/roster.md` if the exact role is not obvious.
+3. Load only the relevant `sub/*/subskill.md` files.
+4. Read `references/strategy/QUICKSTART.md` before any full-pipeline orchestration.
+5. Keep project truth in the repo being edited; Agency roles are execution overlays, not replacement project docs.
+
+## Routing Heuristics
+
+- UI, frontend, components, accessibility, browser UX -> frontend, UI, UX, or whimsy roles
+- APIs, infrastructure, security, data, AI, mobile -> backend, DevOps, security, data, AI, or mobile roles
+- Discovery, prioritization, project planning -> product and project-management roles
+- QA, evidence, performance, API verification, accessibility -> testing roles
+- Growth, content, social, launch, app stores -> marketing roles
+- Ops, finance, compliance, analytics, support -> support roles
+- Spatial/XR/visionOS workflows -> spatial-computing roles
+- Cross-functional orchestration -> `sub/agents-orchestrator/subskill.md` plus NEXUS references
+
+## Key References
+
+- Roster: `references/roster.md`
+- NEXUS quick start: `references/strategy/QUICKSTART.md`
+- Master strategy: `references/strategy/nexus-strategy.md`
+- Phase playbooks: `references/strategy/playbooks/`
+- Scenario runbooks: `references/strategy/runbooks/`
+- Coordination prompts/templates: `references/strategy/coordination/`
+- Worked examples: `references/examples/`
+
+## Minimal Loading Rule
+
+Do not load the whole catalog by default.
+
+Recommended sequence:
+
+1. Read this file.
+2. Read `references/roster.md` only if role selection is unclear.
+3. Load 1-3 role sub-skills from `sub/`.
+4. Pull in NEXUS references only for orchestration-heavy work.
+5. Return to the project repo and execute there.

integrations/codex/agency-agents/agents/openai.yaml

@@ -0,0 +1,4 @@
+interface:
+  display_name: "The Agency"
+  short_description: "Agency Codex meta-skill with NEXUS orchestration"
+  default_prompt: "Use $agency-agents to pick the right Agency specialist or run a NEXUS-style multi-role workflow inside Codex."

integrations/codex/agency-agents/references/examples/README.md

@@ -0,0 +1,48 @@
+# Examples
+
+This directory contains example outputs demonstrating how the agency's agents can be orchestrated together to tackle real-world tasks.
+
+## Why This Exists
+
+The agency-agents repo defines dozens of specialized agents across engineering, design, marketing, product, support, spatial computing, and project management. But agent definitions alone don't show what happens when you **deploy them all at once** on a single mission.
+
+These examples answer the question: *"What does it actually look like when the full agency collaborates?"*
+
+## Contents
+
+### [nexus-spatial-discovery.md](./nexus-spatial-discovery.md)
+
+**What:** A complete product discovery exercise where 8 agents worked in parallel to evaluate a software opportunity and produce a unified plan.
+
+**The scenario:** Web research identified an opportunity at the intersection of AI agent orchestration and spatial computing. The entire agency was then deployed simultaneously to produce:
+
+- Market validation and competitive analysis
+- Technical architecture (8-service system design with full SQL schema)
+- Brand strategy and visual identity
+- Go-to-market and growth plan
+- Customer support operations blueprint
+- UX research plan with personas and journey maps
+- 35-week project execution plan with 65 sprint tickets
+- Spatial interface architecture specification
+
+**Agents used:**
+| Agent | Role |
+|-------|------|
+| Product Trend Researcher | Market validation, competitive landscape |
+| Backend Architect | System architecture, data model, API design |
+| Brand Guardian | Positioning, visual identity, naming |
+| Growth Hacker | GTM strategy, pricing, launch plan |
+| Support Responder | Support tiers, onboarding, community |
+| UX Researcher | Personas, journey maps, design principles |
+| Project Shepherd | Phase plan, sprints, risk register |
+| XR Interface Architect | Spatial UI specification |
+
+**Key takeaway:** All 8 agents ran in parallel and produced coherent, cross-referencing plans without coordination overhead. The output demonstrates the agency's ability to go from "find an opportunity" to "here's the full blueprint" in a single session.
+
+## Adding New Examples
+
+If you run an interesting multi-agent exercise, consider adding it here. Good examples show:
+
+- Multiple agents collaborating on a shared objective
+- The breadth of the agency's capabilities
+- Real-world applicability of the agent definitions