aidevops 3.32.309 → 3.32.311

package/README.md

@@ -9,7 +9,7 @@
 
 Most AI tools still leave you doing the hard coordination yourself: finding the right context, choosing a model, protecting secrets, managing branches, watching CI, spotting stuck work, and remembering what went wrong last time. aidevops puts structure around that work so agents can share context, work safely in parallel, spend model budget where it matters, and leave the system better than they found it.
 
-> **Recommended setup:** [OpenCode](https://opencode.ai/) + OpenAI GPT-5.6 models. aidevops routes Luna to bounded work, Terra to general implementation, and Sol to consequential reasoning and synthesis. Claude models (Anthropic) remain fully supported fallbacks, and other model providers are evaluated as their quality, latency, and cost profiles change.
+> **Recommended setup:** [OpenCode](https://opencode.ai/) + OpenAI GPT-5.6 Luna / Terra and GPT-6 Astra. aidevops routes Luna to bounded work, Terra to general implementation, and Astra at low reasoning to consequential reasoning and synthesis. Claude models (Anthropic) remain fully supported fallbacks, and other model providers are evaluated as their quality, latency, and cost profiles change.
 
 *"Scope a mission to redesign the landing pages — break it into milestones, dispatch workers in parallel, validate each milestone, and track budget across the whole project."*
 
@@ -60,7 +60,7 @@ The result: an AI operations platform that manages projects across every busines
 [![Copyright](https://img.shields.io/badge/Copyright-Marcus%20Quinn%202025--2026-blue.svg)](https://github.com/marcusquinn)
 
 <!-- Release & Version Info -->
-[![Version](https://img.shields.io/badge/Version-3.32.309-blue.svg)](https://github.com/marcusquinn/aidevops/releases)
+[![Version](https://img.shields.io/badge/Version-3.32.311-blue.svg)](https://github.com/marcusquinn/aidevops/releases)
 [![npm version](https://img.shields.io/npm/v/aidevops)](https://www.npmjs.com/package/aidevops)
 [![Homebrew](https://img.shields.io/badge/homebrew-marcusquinn%2Ftap-orange)](https://github.com/marcusquinn/homebrew-tap)
 [![GitHub repository](https://img.shields.io/badge/github-repository-181717.svg?logo=github)](https://github.com/marcusquinn/aidevops)
@@ -92,7 +92,7 @@ the required notices and preferred credit text.
 
 - **Purpose**: AI-assisted DevOps automation framework
 - **Install**: `npm install -g aidevops && aidevops update`
-- **Recommended runtime/models**: OpenCode + OpenAI GPT-5.6 Luna / Terra / Sol
+- **Recommended runtime/models**: OpenCode + OpenAI GPT-5.6 Luna / Terra / GPT-6 Astra
 - **Entry**: `aidevops` CLI, `~/.aidevops/agents/AGENTS.md`
 - **Stack**: Bash scripts, TypeScript (Bun), MCP servers
 - **Recent focus**: evidence-based model routing, indexed worker worktrees, audited deployment copies, release provenance, and OpenCode control-plane safety
@@ -119,7 +119,7 @@ the required notices and preferred credit text.
 - `/auto-browse` - Learn, optimize, and graduate repeatable browser operations and web data-mining workflows
 - `/report-render` - Render report-ready Markdown or JSON to HTML with sticky TOC, print CSS, evidence badges, and source cards for PDF export
 - `/report-token-use` - Generate a local per-session token, model, compaction, and MCP-use report
-- `/optimize-tiers` - Compare production routing telemetry with sealed historical model replay before changing model defaults
+- `/optimize-tiers` - Optionally investigate unresolved model comparisons with production telemetry and sealed historical replay
 - `/pulse` - Run the autonomous supervisor loop for dispatch, merge, diagnostics, and stuck-work recovery
 - `/serve-sim` / `serve-sim-helper.sh` - Exercise mobile web flows in simulator-backed local previews
 
@@ -287,7 +287,7 @@ git clone https://github.com/marcusquinn/aidevops.git ~/Git/aidevops
 - Ensure all PATH and alias changes work in both bash, zsh, and fish
 - When Claude Code is installed, add a `claude` alias that runs `claude --dangerously-skip-permissions` (skips per-tool permission prompts). Re-running setup updates the alias automatically. To grant permissions per-session instead, press **Shift-Tab** inside Claude Code to cycle through permission modes (default → skip permissions → auto-approve).
 
-**New users: Start [OpenCode](https://opencode.ai/) and type `/onboarding`** to configure your services interactively. OpenCode is the recommended tool for aidevops; the default routing uses OpenAI GPT-5.6 Luna, Terra, and Sol across the simple, standard, and thinking tiers. The onboarding wizard will:
+**New users: Start [OpenCode](https://opencode.ai/) and type `/onboarding`** to configure your services interactively. OpenCode is the recommended tool for aidevops; the default routing uses OpenAI GPT-5.6 Luna, Terra, and GPT-6 Astra across the simple, standard, and thinking tiers. The onboarding wizard will:
 - Explain what **[aidevops](https://aidevops.sh)** can do
 - Ask about your work to give personalized recommendations
 - Show which services are configured vs need setup
@@ -446,7 +446,9 @@ See `.agents/tools/task-management/beads.md` for complete documentation and inst
 
 ### OpenAI Models in OpenCode (Recommended)
 
-OpenCode with OpenAI is the current recommended aidevops setup. The default routing uses GPT-5.6 Luna for bounded low-consequence work, Terra for established-pattern implementation, and Sol for consequential decisions, architecture, and synthesis-heavy work.
+OpenCode with OpenAI is the current recommended aidevops setup. Delegation defaults are GPT-5.6 Luna **medium** for bounded low-consequence work, Terra **high** for established-pattern implementation, and GPT-6 Astra **low** for consequential decisions, architecture, and synthesis-heavy work. Headless thinking workers escalate Astra reasoning **low → medium → high** only on an evidence-bearing capability-limit signal; operational failures never trigger that escalation. Interactive children retain their parent's reasoning ceiling. See [model routing](.agents/tools/context/model-routing.md) for overrides and escalation boundaries.
+
+Optimise through ordinary verified work, retries and parent repair rather than a prerequisite benchmark project. API prices are directional resource-cost estimates, not ChatGPT subscription allowance percentages. See [practical model and effort selection](.agents/tools/context/model-routing.md#practical-model-and-effort-selection) for the strategy, estimate scope and safe escalation rules.
 
 **Authenticate via the pool:**
 
@@ -457,8 +459,8 @@ aidevops model-accounts-pool add openai
 
 **Why this is the default:**
 
-- **Best current cross-tier results** — strongest observed balance across interactive Build+, workers, review, and dispatch tiers
-- **Tiered cost/latency split** — Luna for bounded work, Terra for general implementation, and Sol where deeper judgement is worth the extra budget
+- **Pragmatic defaults** — reversible choices improved through actual completion and repair evidence, not a claim of benchmark superiority
+- **Tiered cost/latency split** — Luna for bounded work, Terra for general implementation, and Astra where deeper judgement is worth the extra budget
 - **Provider isolation** — OpenAI accounts rotate independently from Anthropic, Google, Cursor, and local providers
 - **Fallback-friendly** — Claude, Gemini, Cursor, and local models remain available when a task or rate-limit profile calls for them
 
@@ -576,7 +578,7 @@ The secure workflow is included at `.github/workflows/opencode-agent.yml`.
 
 See `.agents/tools/git/opencode-github-security.md` for the full security documentation.
 
-**Supported AI tool:** [OpenCode](https://opencode.ai/) is the recommended and tested AI coding tool for aidevops. All features, agents, and workflows are designed and tested for OpenCode first. The default OpenAI routing maps GPT-5.6 Luna, Terra, and Sol to the simple, standard, and thinking tiers. [Claude](https://claude.ai/) models (Anthropic) remain fully supported fallbacks, and other providers are tested as their capabilities change.
+**Supported AI tool:** [OpenCode](https://opencode.ai/) is the recommended and tested AI coding tool for aidevops. All features, agents, and workflows are designed and tested for OpenCode first. The default OpenAI routing maps GPT-5.6 Luna, Terra, and GPT-6 Astra to the simple, standard, and thinking tiers. [Claude](https://claude.ai/) models (Anthropic) remain fully supported fallbacks, and other providers are tested as their capabilities change.
 
 The native `ai_research` tool uses OpenCode's configured providers and canonical `simple`, `standard`, and `thinking` workload tiers; it does not require a specific provider credential.
 
@@ -584,7 +586,7 @@ The native `ai_research` tool uses OpenCode's configured providers and canonical
 
 - **[OpenCode](https://opencode.ai/)** - The recommended AI coding agent. Powerful agentic TUI/CLI with native MCP support, Tab-based agent switching, LSP integration, plugin ecosystem, and excellent DX. All aidevops features are designed and tested for OpenCode first.
 - **[OpenCode Zen](https://opencode.ai/)** - Free tier of OpenCode with included models. Start working with AI straight away at no cost -- no API keys or subscriptions required.
-- **OpenAI GPT-5.6 Luna / Terra / Sol** - Recommended tier family for aidevops today: Luna for bounded work, Terra for routine implementation, and Sol for consequential reasoning and high-impact decisions.
+- **OpenAI GPT-5.6 Luna / Terra / GPT-6 Astra** - Recommended tier mapping for aidevops today: Luna for bounded work, Terra for routine implementation, and Astra for consequential reasoning and high-impact decisions.
 - **[Claude](https://claude.ai/)** (Anthropic) - Fully supported alternative provider. Claude models remain useful for fallback, cross-provider verification, and users with Claude Pro/Max OAuth access.
 - **[Tabby](https://tabby.sh/)** - Recommended terminal. Colour-coded Profiles per project/repo, **auto-syncs tab titles with git/session context and marks OpenCode turns from the first submitted message as ⚪, 🔴, 🟡, or 🟢.**
 - **[Zed](https://zed.dev/)** - Recommended editor. High-performance with AI integration (use with the OpenCode Agent Extension).
@@ -871,7 +873,7 @@ Supervisor (pulse loop)
 │   ├── task_assignment → worker inbox
 │   ├── status_report → coordinator outbox
 │   └── broadcast → all agents
-└── Model Routing (tier-based: GPT-5.6 Luna / Terra / Sol / provider fallbacks)
+└── Model Routing (tier-based: GPT-5.6 Luna / Terra / GPT-6 Astra / provider fallbacks)
 ```
 
 **Key components:**
@@ -1365,7 +1367,6 @@ The setup script offers to install these tools automatically.
 
 ### **AI Prompt Optimization**
 
-- **[Augment Context Engine](https://docs.augmentcode.com/context-services/mcp/overview)**: Semantic codebase retrieval with deep code understanding
 - **[Repomix](https://repomix.com/)**: Pack codebases into AI-friendly context (80% token reduction with compress mode)
 - **[DSPy](https://dspy.ai/)**: Framework for programming with language models
 - **[DSPyGround](https://dspyground.com/)**: Interactive playground for prompt optimization
@@ -1450,13 +1451,12 @@ See `.agents/tools/ocr/glm-ocr.md` for batch processing, PDF workflows, and Peek
 
 **Model Context Protocol servers for real-time AI assistant integration.** The framework configures these MCPs for **[OpenCode](https://opencode.ai/)** (TUI, Desktop, and Extension for Zed/VSCode).
 
-### **All Supported MCPs (20 available)**
+### **All Supported MCPs (19 available)**
 
 MCP integrations use reviewed local package runners or remote HTTPS endpoints. Security-sensitive local integrations use exact package pins where required; setup may globally cache selected tools for faster startup. Run `setup.sh` or `aidevops update-tools` to refresh managed tooling.
 
 | MCP | Purpose | Tier | API Key Required |
 |-----|---------|------|------------------|
-| [Augment Context Engine](https://docs.augmentcode.com/context-services/mcp/overview) | Semantic codebase retrieval | Global | Yes (Augment account) |
 | [Claude Code MCP](https://github.com/steipete/claude-code-mcp) | Claude as sub-agent | Global | No |
 | [Amazon Order History](https://github.com/marcusquinn/amazon-order-history-csv-download-mcp) | Order data extraction | Per-agent | No |
 | [Chrome DevTools](https://chromedevtools.github.io/devtools-protocol/) | Browser debugging & automation | Per-agent | No |
@@ -1503,7 +1503,6 @@ These use direct API calls via curl, avoiding MCP server startup entirely:
 
 **Context & Codebase:**
 
-- [Augment Context Engine](https://docs.augmentcode.com/context-services/mcp/overview) - Semantic codebase retrieval with deep code understanding
 - [llm-tldr](https://github.com/parcadei/llm-tldr) - Semantic code analysis with 95% token savings ([details below](#llm-tldr---semantic-code-analysis))
 - [Context7](https://context7.com/) - Real-time documentation access for thousands of libraries
 - [Repomix](https://github.com/yamadashy/repomix) - Pack codebases into AI-friendly context
@@ -1848,84 +1847,6 @@ Repomix runs as an MCP server for direct AI assistant integration:
 
 See `.agents/tools/context/context-builder.md` for complete documentation.
 
-## **Augment Context Engine - Semantic Codebase Search**
-
-[Augment Context Engine](https://docs.augmentcode.com/context-services/mcp/overview) provides semantic codebase retrieval - understanding your code at a deeper level than simple text search. It's the recommended tool for real-time interactive coding sessions.
-
-### Why Augment Context Engine?
-
-| Feature | grep/glob | Augment Context Engine |
-|---------|-----------|------------------------|
-| Text matching | Exact patterns | Semantic understanding |
-| Cross-file context | Manual | Automatic |
-| Code relationships | None | Understands dependencies |
-| Natural language | No | Yes |
-
-Use it to:
-
-- Find related code across your entire codebase
-- Understand project architecture quickly
-- Discover patterns and implementations
-- Get context-aware code suggestions
-
-### Quick Setup
-
-```bash
-# 1. Install Auggie CLI (requires Node.js 22+)
-npm install -g @augmentcode/auggie@prerelease
-
-# 2. Authenticate (opens browser)
-auggie login
-
-# 3. Verify installation
-auggie token print
-```
-
-### MCP Integration
-
-Add to your AI assistant's MCP configuration:
-
-**OpenCode** (`~/.config/opencode/opencode.json`):
-
-```json
-{
-  "mcp": {
-    "augment-context-engine": {
-      "type": "local",
-      "command": ["auggie", "--mcp"],
-      "enabled": true
-    }
-  }
-}
-```
-
-**claude-code CLI**:
-
-```bash
-claude mcp add-json auggie-mcp --scope user '{"type":"stdio","command":"auggie","args":["--mcp"]}'
-```
-
-### Verification
-
-Test with this prompt:
-
-```text
-What is this project? Please use codebase retrieval tool to get the answer.
-```
-
-The AI should provide a semantic understanding of your project architecture.
-
-### Repomix vs Augment Context Engine
-
-| Use Case | Tool | When to Use |
-|----------|------|-------------|
-| **Interactive coding** | Augment Context Engine | Real-time semantic search during development |
-| **Share with external AI** | Repomix | Self-contained snapshot for ChatGPT, Claude web, etc. |
-| **Architecture review** | Repomix (compress) | 80% token reduction, structure only |
-| **CI/CD integration** | Repomix GitHub Action | Automated context in releases |
-
-See `.agents/tools/context/augment-context-engine.md` for complete documentation.
-
 ### llm-tldr - Semantic Code Analysis
 
 [llm-tldr](https://github.com/parcadei/llm-tldr) extracts code structure and semantics, saving ~95% tokens compared to raw code. From the [Continuous-Claude](https://github.com/parcadei/Continuous-Claude-v3) project.
@@ -2065,6 +1986,23 @@ Primary agents live at `.agents/<name>.md`. Each is a domain expert with its own
 
 #### How to invoke a main agent
 
+Launch terminal clients directly with Build+ selected by default, or name a main agent:
+
+```bash
+aidevops codex
+aidevops codex automate
+aidevops claude seo
+aidevops opencode research
+```
+
+These shortcuts apply each CLI's supported full-access options for the session.
+They also support Cursor Agent, Gemini, Continue, Kilo, Kiro CLI, Aider, Amp, Kimi,
+Qwen and Droid (which retains its configured interactive autonomy). Use
+`aidevops codex --list-agents` to list main agents, `--dry-run` to preview a launch,
+and `--native` for native subcommands without injected defaults. See
+[interactive runtime launchers](.agents/tools/ai-assistants/runtime-launchers.md)
+for aliases, argument forwarding and runtime limitations.
+
 | Client | Invocation |
 |---|---|
 | **OpenCode** | Tab through the agent picker in the UI — Build+ is the default. Main agents are registered as top-level agents in OpenCode's config. |
@@ -2073,6 +2011,12 @@ Primary agents live at `.agents/<name>.md`. Each is a domain expert with its own
 
 The `aidevops-` prefix differentiates framework commands from each client's native slash commands and groups them alphabetically in the command picker. It applies to **every** aidevops slash command — not just main agents — so `/aidevops-preflight`, `/aidevops-release`, `/aidevops-commit` etc. all sort together in your `/` menu.
 
+Codex setup/update also installs global `AGENTS.md` guidance, native workflow
+skills and compatible lifecycle hooks. Build+ loads by default; use `/skills`
+for workflows and `/hooks` to review/trust new hooks. Existing personal settings
+are preserved. For full-access launch flags and runtime differences, see
+[Codex CLI integration](.agents/tools/ai-assistants/codex-cli.md).
+
 #### Supported AI clients (14)
 
 The framework installs itself across these clients. Slash commands, agent definitions, and (optionally) session-memory mining are wired up per-client by `setup.sh`, gated on per-client feature flags in `.agents/scripts/runtime-registry.sh`.
@@ -2081,7 +2025,7 @@ The framework installs itself across these clients. Slash commands, agent defini
 |---|---|---|---|---|
 | OpenCode | ✅ `~/.config/opencode/command/` | config-based | ✅ default | Native tab-through primary agents |
 | Claude Code | ✅ `~/.claude/commands/` | ✅ `~/.claude/agents/` | ✅ default | Full feature parity |
-| Codex CLI | ✅ `~/.codex/prompts/` | — | ✅ default | Invoked as `/prompts:aidevops-<name>` |
+| Codex CLI | ✅ native skills + legacy `~/.codex/prompts/` | Workflow skills | ✅ default | `$aidevops-<name>`; legacy `/prompts:aidevops-<name>` |
 | Cursor | ✅ `~/.cursor/commands/` (≥1.6) | ✅ `~/.cursor/agents/` | ✅ default | Frontmatter stripped (not supported) |
 | Droid (Factory) | ✅ `~/.factory/commands/` | — | opt-in | — |
 | Gemini CLI | ✅ `~/.gemini/commands/` | — | opt-in | Converted to TOML (`prompt = """..."""`) |

package/VERSION

@@ -1 +1 @@
-3.32.309
+3.32.311

package/aidevops.sh

@@ -5,7 +5,7 @@
 # AI DevOps Framework CLI
 # Usage: aidevops <command> [options]
 #
-# Version: 3.32.309
+# Version: 3.32.311
 
 set -euo pipefail
 
@@ -1123,7 +1123,10 @@ _help_commands() {
 	echo "  gpt56-context <cmd> Manage the 300K GPT-5.6 OpenCode context cap (enable/disable/status)"
 	echo "  client-format      Client request format alignment (extract/check/canary/monitor)"
 	echo "  opencode-db <cmd>  OpenCode SQLite maintenance/session lookup (check/report/sessions/maintain/window/status/install)"
-	echo "  opencode [args]    Launch OpenCode with aidevops per-session DB isolation"
+	echo "  opencode [agent] [args]    Launch OpenCode with Build+ and per-session DB isolation"
+	echo "  codex|claude [agent] [args]  Launch with full permissions (default agent: build-plus)"
+	echo "  <runtime> [agent] [args]     Also cursor, droid, gemini, continue, kilo, kiro, aider, amp, kimi, qwen"
+	echo "                              Use <runtime> --help or --list-agents for launch options"
 	echo "  opencode-desktop   Launch/install OpenCode Desktop with aidevops DB isolation"
 	echo "  opencode-sandbox   Test OpenCode versions in isolation (install/run/check/clean)"
 	echo "  approve <cmd>      Cryptographic issue/PR approval (setup/issue/pr/verify/status)"
@@ -1975,7 +1978,8 @@ main() {
 	client-format) _cmd_client_format "$@" ;;
 	github-app-auth | github-app | gh-auth) _dispatch_helper "github-app-auth-helper.sh" "github-app-auth-helper.sh" "$@" ;;
 	opencode-db | oc-db) _dispatch_helper "opencode-db-maintenance-helper.sh" "opencode-db-maintenance-helper.sh" "$@" ;;
-	opencode | oc) _dispatch_helper "opencode-launcher-helper.sh" "opencode-launcher-helper.sh" "$@" ;;
+	opencode | oc | codex | claude | claude-code | cursor | cursor-agent | droid | gemini | gemini-cli | continue | cn | kilo | kiro | kiro-cli | aider | amp | kimi | qwen | windsurf)
+		_dispatch_helper "runtime-launcher-helper.sh" "runtime-launcher-helper.sh" "$command" "$@" ;;
 	opencode-desktop | oc-desktop) _dispatch_helper "opencode-launcher-helper.sh" "opencode-launcher-helper.sh" desktop "$@" ;;
 	opencode-sandbox | oc-sandbox) _dispatch_helper "opencode-sandbox-helper.sh" "opencode-sandbox-helper.sh" "$@" ;;
 	review-gate | review_gate) _dispatch_helper "review-gate-config-helper.sh" "review-gate-config-helper.sh" "$@" ;;

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "aidevops",
-  "version": "3.32.309",
+  "version": "3.32.311",
   "description": "AI DevOps Framework - AI-assisted development workflows, code quality, and deployment automation",
   "type": "module",
   "packageManager": "bun@1.3.11",

package/setup.sh

@@ -17,7 +17,7 @@ fi
 # AI Assistant Server Access Framework Setup Script
 # Helps developers set up the framework for their infrastructure
 #
-# Version: 3.32.309
+# Version: 3.32.311
 #
 # Quick Install:
 #   npm install -g aidevops && aidevops update          (recommended)