agents-united 0.9.0 → 0.10.0

package/README.md

@@ -19,7 +19,7 @@ Curated teams of orchestrators, sub-agents, skills, and workflows — installed
 - **🌐 One Library, Every Assistant**: Author in `.agents/` as your single source of truth. Agents United automatically projects and translates compatible configurations to **Google Antigravity 2.0 / Gemini**, **Anthropic Claude Code**, **Cursor**, **Cline**, **OpenCode**, and **Codex / AGENTS.md**.
 - **🚀 Cline Native Activation**: Bundles activate **automatically** in any Cline CLI session — skills discovered natively from `.agents/skills/`, configured-agent roles (`.cline/agents/*.yml`) exposed as spawnable `subagent_*` tools, coordinator rules (`.cline/rules/`), slash-command workflows (`.cline/workflows/`), and spec-conformant Agent Plugin packages (`plugin.json`, agent-plugins.org) in `.agents/plugins/<bundle>/`. No install step required; `agents start` remains available as an optional pre-seeded team-session launcher.
 - **🏛️ 8 Department Domains & 26 Bundles**: Complete coverage across Software Engineering, System Architecture, Product Design, Growth & Marketing, Security, Deep Research, Business Strategy, and Universal Meta-Skills.
-- **🤖 58 Specialized Agents, 91 Skills & 69 Workflows**: 8 Lead Orchestrators coordinating 50 domain sub-agents, backed by 91 production-grade runbooks and 69 deterministic workflows.
+- **🤖 59 Specialized Agents, 91 Skills & 69 Workflows**: 9 Lead Orchestrators coordinating 50 domain sub-agents, backed by 91 production-grade runbooks and 69 deterministic workflows.
 - **⚡ Modern Cloud & AI Tooling**: First-class runbooks for Modal.com, Replicate, RunPod, local LLMs/vLLM, LangChain, LlamaIndex, Qdrant, Vercel, Supabase, Turso, and Azure Bicep.
 - **🏢 Organization Bundles (Tier 2 / Experimental)**: Cross-functional multi-disciplinary teams with Tri-Tier MCP execution modes and smart auto-remediation (`digital-agency`).
 - **🛡️ Built-in Zero-Trust Safety & Git Guardrails**: Hard-coded branch protection (`main`/`master`), zero force-push policy, secret redaction, serverless GPU cost ceilings, and PII scrubbing.
@@ -119,7 +119,13 @@ Agents United provides **24 curated bundles**, **46 specialized agents** (8 Orch
 | **🔒 Security Operations** | `security-operations` | — | `orchestrator-security` |
 | **🔬 Deep Technical Research** | `deep-research` | — | `orchestrator-research` |
 | **💼 Business Strategy & Economics** | `business-strategy` | — | `orchestrator-business` |
-| **🏢 Organization Bundles (Experimental)** | `digital-agency` ⚡ | *(Cross-functional composite)* | `orchestrator-marketing` (Campaign Director / Chris) |
+| **🏢 Organization Bundles (Experimental)** | `digital-agency` ⚡ | *(Cross-functional composite)* | `orchestrator-digital-agency` (Campaign Director / Chris) |
+
+> [!NOTE]
+> **Bundle Development Status**: Currently, `software-engineering` and `digital-agency` are actively implemented and production-verified. Except for these two bundles, all other domain bundles are currently under active development.
+
+> [!IMPORTANT]
+> **Tested Environments & Community Feedback**: Agents United has been manually tested and verified on **Cline** (CLI & VS Code Extension), **Google Antigravity Desktop App**, and **Google Antigravity CLI**. We are actively looking for user feedback, issue reports, and telemetry for other environments (Claude Code, Cursor, OpenCode, Codex).
 
 <details>
 <summary><strong>🔍 Click to expand full Department & Agent Roster Breakdown</strong></summary>
@@ -173,7 +179,7 @@ Agents United provides **24 curated bundles**, **46 specialized agents** (8 Orch
 - **Essentials Base (`business-strategy`)**: `subagent-business-panel-experts`
 
 #### 🏢 Organization Bundles (Tier 2 / Experimental)
-- **Digital Product Agency (`digital-agency` ⚡)**: 10-agent cross-functional AstrolabsAI team (Chris / Ava / Kaan / Jamileh / Yavuz / Jale + engineering, QA & compliance) with Tri-Tier MCP execution modes and the Subagent-First Planning Dialogue Loop (ADR 0014).
+- **Digital Product Agency (`digital-agency` ⚡)**: 10-agent cross-functional AstrolabsAI team (Chris / Ava / Kaan / Jamileh / Yavuz / Jale + frontend architect, SEO specialist, QA automation lead, and compliance GRC specialist) with Tri-Tier MCP execution modes and the Subagent-First Planning Dialogue Loop (ADR 0014).
 
 </details>
 
@@ -408,7 +414,7 @@ Organization bundles operate across three dynamic execution envelopes:
 
 | Mode | Capability Envelope | In-Session Behavior |
 | :--- | :--- | :--- |
-| **🚀 Fully Operational** | Authenticated live MCP tools (`github`, `firecrawl`, `context7`, `playwright`, `stitch`, `figma`) | Subagents execute live browser automation, deep web crawling, design token extraction, and automated GitHub PRs. |
+| **🚀 Fully Operational** | Authenticated live MCP tools (`github`, `firecrawl`, `context7`, `playwright`, `markitdown`, `chrome-devtools-mcp`, `stitch`, `figma`) | Subagents execute live browser automation, deep web crawling, design token extraction, and automated GitHub PRs. |
 | **🌿 Limited Operational** | Unauthenticated MCP tools + Native CLI Fallbacks (`git`, `curl`, code generators) | Executes immediate real-world work using local browsers and terminal tools without requiring API keys or cloud tokens. |
 | **💡 Brainstorming Mode** | Pure architectural ideation, strategy, copywriting & specification drafting | Zero tool-calling overhead; ideal for offline or air-gapped environments. |
 
@@ -428,6 +434,28 @@ Organization bundles can opt into a **delegation-first planning protocol**: the
 
 Enabled for `digital-agency`. See [ADR 0014](./docs/adr/0014-subagent-first-planning-loop.md) and [Plan 012](./plans/012-subagent-first-planning-loop.md).
 
+### 4-Tier Agency Assembly Line (Deterministic Execution DAG)
+
+The `digital-agency` bundle executes through a structured 4-tier delivery pipeline ensuring that technical and creative assets are rigorously cross-pollinated:
+
+```text
+Ava (Growth / LTV:CAC) + Chris (Director)
+  │ (Audience brief, target economics, core offer)
+  ▼
+Kaan (Direct Response CRO) + Jamileh (Design System) + Yavuz (SEO Content)
+  │ (Design tokens, persuasive hooks, article pillars, copy schemas)
+  ▼
+Frontend Architect (React / Tailwind) + SEO Specialist + Jale (Social / Ads)
+  │ (Production components, data-testid, GTM dataLayer, ad carousels, UTM taxonomy)
+  ▼
+QA Automation Lead (Playwright E2E) + Compliance GRC Specialist (FTC / GDPR / CASL)
+  │ (100% green test assertions, CMP gating verification, disclosure compliance)
+  ▼
+🚀 Production-Ready Campaign Artifacts & Verified PR
+```
+
+- **Cross-Host Capability Projection**: Google Antigravity features are seamlessly mapped to Cline CLI (ADR 0013) — mapping tool primitives (`view_file` → `read_file`, `replace_file_content` → `replace_in_file`, `run_command` → `execute_command`), converting frontmatter rules to native `.cline/rules/`, capping subagent iterations (`maxIterations: 8`), and preserving `@path/to/file` multimodal asset ingestion and KaTeX econometric calculations across both platforms.
+
 ### Planner-Orchestrator Mode (ADR 0015)
 
 **Tier-1 Domain Bundles** operate in **Planner-Orchestrator Mode**: the orchestrator plans **solo** with the user — Socratic alignment via `/grill-me` / `/grill-with-docs`, direct skill consultation, no specialist spawns during planning. A **Planning Aid Boundary** allows provisional estimates during planning but defers concrete deliverables to the delegation map. Execution is then delegated to `subagent_*` tools via a solo-composed **Delegation Map**. This mode is the default for all 30 domain bundles (engineering, design, security, business, research, architecture, marketing); `digital-agency` and future Organization Bundles remain in Subagent-First Mode. See [ADR 0015](./docs/adr/0015-planner-orchestrator-mode-for-domain-bundles.md) and [Plan 013](./plans/013-planner-orchestrator-mode-for-domain-bundles.md).
@@ -482,6 +510,9 @@ All projections are recorded in `agents-united.json` under `projectedTo`, so `re
 
 > ⚠️ **Per-runtime caveats.** Antigravity-only frontmatter keys — `hooks:`, `permissionMode:`, `commandExecutionPolicy:`, `mainAgent:`, `subagent:`, and `type:` — **do not execute** in other runtimes. Projected orchestrator agents degrade gracefully to "system prompt + tool list" subagents; behavior is documented, not faked. Cross-agent `invoke_subagent` orchestration works only in Antigravity and is dropped elsewhere with a warning. To add more runtimes later, re-run `agents add <bundle> -t agents --fanout <hosts>`.
 
+> [!IMPORTANT]
+> **Host Testing & Community Feedback**: Agents United has been tested manually on **Cline**, the **Google Antigravity Desktop App**, and the **Antigravity CLI**. Projections for other environments (Anthropic Claude Code, Cursor, OpenCode, Codex / AGENTS.md) are generated according to their platform specifications. We are actively looking for user feedback across other CLIs and runtimes!
+
 ---
 
 ## 🛡️ Built-in Safety & Git Guardrails

package/dist/cli.js

@@ -3612,7 +3612,7 @@ var BUNDLE_DISPLAY_NAMES = {
     summary: "Market analysts, monetization experts, and executive spec panels"
   },
   "digital-agency": {
-    title: "Digital Agency (TBA soon)",
+    title: "Digital Agency",
     summary: "Full-service digital product agency with web dev, mobile, design, SEO, and growth marketing"
   },
   "universal-orchestration": {