pantheon-opencode 1.1.1-beta.9 → 1.1.3-beta.0

package/src/routing.yml

@@ -2,7 +2,6 @@ version: 1.0
 agents:
   zeus:
     role: Central orchestrator
-    model_tier: code
     reasoning_effort: medium
     user_invocable: true
     can_delegate: true
@@ -41,7 +40,6 @@ agents:
     - incremental-implementation
   athena:
     role: Strategic planner
-    model_tier: premium
     reasoning_effort: high
     user_invocable: true
     can_delegate: true
@@ -61,7 +59,6 @@ agents:
     - memory-bank
   apollo:
     role: Investigation scout
-    model_tier: free
     reasoning_effort: low
     user_invocable: false
     can_delegate: false
@@ -79,7 +76,6 @@ agents:
     - auto-continue
   hermes:
     role: Backend specialist
-    model_tier: code
     reasoning_effort: medium
     user_invocable: true
     can_delegate: true
@@ -92,6 +88,10 @@ agents:
     - obsolete_lib_detection
     - memory_store
     - pantheon_code_mode
+    subagent_can_delegate_to:
+    - apollo
+    - themis
+    - mnemosyne
     skills:
     - tdd-with-agents
     - file-prompts
@@ -100,7 +100,6 @@ agents:
     - incremental-implementation
   aphrodite:
     role: Frontend specialist
-    model_tier: code
     reasoning_effort: medium
     user_invocable: true
     can_delegate: false
@@ -113,13 +112,13 @@ agents:
     - deprecated_npm_detection
     - memory_store
     - pantheon_resources_memory_bank
+    - pantheon_code_mode
     skills:
     - tdd-with-agents
     - visual-review-pipeline
     - file-prompts
   demeter:
     role: Database specialist
-    model_tier: code
     reasoning_effort: medium
     user_invocable: true
     can_delegate: false
@@ -132,12 +131,12 @@ agents:
     - index_strategy
     - memory_store
     - memory_recall
+    - pantheon_code_mode
     skills:
     - tdd-with-agents
     - incremental-implementation
   themis:
     role: Quality gate
-    model_tier: premium
     reasoning_effort: high
     user_invocable: true
     can_delegate: true
@@ -151,6 +150,7 @@ agents:
     - deprecation_audit
     - memory_search
     - pantheon_resources
+    - pantheon_code_mode
     subagent_can_delegate_to:
     - mnemosyne
     skills:
@@ -160,7 +160,6 @@ agents:
     - context-compression
   prometheus:
     role: Infrastructure specialist
-    model_tier: code
     reasoning_effort: medium
     user_invocable: true
     can_delegate: false
@@ -178,7 +177,6 @@ agents:
     - incremental-implementation
   hephaestus:
     role: AI pipelines
-    model_tier: code
     reasoning_effort: medium
     user_invocable: true
     can_delegate: false
@@ -190,12 +188,12 @@ agents:
     - embedding_strategy
     - ai_workflow_composition
     - memory_search
+    - pantheon_code_mode
     skills:
     - tdd-with-agents
     - auto-continue
   nyx:
     role: Observability
-    model_tier: free
     reasoning_effort: low
     user_invocable: true
     can_delegate: false
@@ -207,12 +205,12 @@ agents:
     - alerting
     - monitoring_dashboards
     - pantheon_resources_routing
+    - pantheon_code_mode
     skills:
     - security-hardening
     - auto-continue
   gaia:
     role: Remote sensing domain specialist
-    model_tier: free
     reasoning_effort: low
     user_invocable: false
     can_delegate: false
@@ -226,7 +224,6 @@ agents:
     - auto-continue
   iris:
     role: GitHub operations
-    model_tier: free
     reasoning_effort: low
     user_invocable: true
     can_delegate: false
@@ -238,17 +235,14 @@ agents:
     - release_management
     - changelog_generation
     - pantheon_resources
-    subagent_can_delegate_to:
-    - mnemosyne
     skills:
     - git-workflow-and-versioning
     - artifact-management
   mnemosyne:
     role: Memory bank
-    model_tier: free
     reasoning_effort: low
     user_invocable: true
-    can_delegate: true
+    can_delegate: false
     can_receive_delegation: true
     capabilities:
     - memory_bank_management
@@ -260,6 +254,7 @@ agents:
     - memory_store
     - memory_recall
     - pantheon_resources_memory_bank
+    - pantheon_code_mode
     subagent_can_delegate_to: []
     skills:
     - artifact-management
@@ -268,7 +263,6 @@ agents:
     - session-goal
   talos:
     role: Hotfix agent
-    model_tier: free
     reasoning_effort: low
     user_invocable: true
     can_delegate: false
@@ -379,19 +373,6 @@ platforms:
     - bash_execution
     - web_fetch
     - web_search
-fallback_chains:
-  premium_agents:
-    primary: deepseek/deepseek-v4-pro
-    secondary: mimo/v2.5
-    emergency: deepseek/deepseek-v4-pro
-  code_agents:
-    primary: deepseek/deepseek-v4-flash
-    secondary: mimo/v2.5
-    emergency: deepseek/deepseek-v4-flash
-  free_agents:
-    primary: deepseek/deepseek-v4-flash
-    secondary: mimo/v2.5
-    emergency: deepseek/deepseek-v4-flash
 background_job_board:
   max_concurrent_per_agent: 3
   signal_dir: .pantheon/deepwork/board-signals/

package/commands/pantheon-focus.md

@@ -1,9 +0,0 @@
----
-description: "Pin a session goal to prevent scope creep. Usage: /pantheon-focus"
-agent: "zeus"
----
-# /pantheon-focus — Session Goal
-
-**What:** Pins a session objective. When active, all agents verify every action against the pinned goal. Off-topic work is flagged and blocked.
-**Usage:** `/pantheon-focus <short objective>`
-**Commands:** `/pantheon-focus list` — show all active goals, `/pantheon-focus clear` — remove all

package/commands/pantheon-forget.md

@@ -1,58 +0,0 @@
----
-description: "Compress old memory entries into summarized form. /pantheon-forget lists entries older than N days, then replaces them with a compressed summary. --auto applies without confirmation. Usage: /pantheon-forget"
-agent: "zeus"
----
-# /pantheon-forget — Memory Range Compression
-
-**What:** Applies range compression to old memory entries — groups entries by time period and namespace, generates a concise summary via fast LLM, replaces N raw entries with 1 summary. Companion to memory_consolidate (merge) vs memory_forget (compress).
-
-## Usage
-
-```
-/pantheon-forget                                    → Lista candidatas (7+ dias, modo interativo)
-/pantheon-forget --days=30                          → Só entradas com 30+ dias
-/pantheon-forget --namespace=sessions               → Namespace específico
-/pantheon-forget --namespace=sessions --days=14     Combinado
-/pantheon-forget --auto                             → Comprime sem confirmação
-/pantheon-forget --auto --days=30 --compress         → Agressivo: 30+ dias + comprimir em 1 arquivo
-/pantheon-forget --dry-run                          → Preview sem modificar
-```
-
-## Flags
-
-| Flag | Default | Descrição |
-|------|---------|-----------|
-| `--days=<N>` | `7` | Idade mínima em dias |
-| `--namespace=<ns>` | `all` | Namespace específico |
-| `--auto` | off | Comprimir sem confirmação |
-| `--dry-run` | off | Preview sem modificar |
-| `--compress` | off | Comprimir tudo em 1 único entry (modo agressivo) |
-
-## How It Works
-
-1. `memory_list()` → obtém entradas, filtra por idade
-2. Agrupa por namespace + mês
-3. Cada grupo: envia para LLM fast com prompt:
-   ```
-   Summarize these N memory entries into 2-3 sentences:
-   [entries...]
-   ```
-4. Substitui grupo por entry único com key `compressed:<namespace>:<YYYY-MM>`
-5. Entradas originais são deletadas via `memory_forget()`
-
-## Output
-
-```
-🗜️ Memory Range Compression
-
-Group: sessions · 2026-06 (15 entries · 1.2K tokens)
-  → "Sessões de junho focadas em migração FastAPI, ajustes de rota e correção de testes."
-  → Compressed: 15 → 1 entry ✅
-
-Group: decisions · 2026-05 (8 entries · 640 tokens)
-  → "Decisões de maio sobre arquitetura: monorepo mantido, SQLAlchemy 2.0 como padrão."
-  → Compressed: 8 → 1 entry ✅
-
-Summary: 2 groups compressed, 23 entries → 2 entries (-91%)
-Tokens saved: ~1.8K
-```

package/commands/pantheon-reflect.md

@@ -1,16 +0,0 @@
----
-description: "Reflect on workflow friction and create improvement action items"
-agent: zeus
----
-# /pantheon-reflect — Workflow Reflection
-
-**What:** After completing a task, reflect on what could be improved. Turns repeated friction into reusable skills, agents, or config improvements.
-**Usage:** `/pantheon-reflect [session-or-task]`
-**Returns:** Structured reflection with action items
-
-## When to Use
-- After completing a complex feature
-- When you notice the same manual steps repeating
-- At end of work session
-
-Delegates to @mnemosyne with `skill: reflect`.

package/commands/pantheon-remember.md

@@ -1,44 +0,0 @@
----
-description: "Store or recall a memory interactively. /pantheon-remember <text> → stores it. /pantheon-remember recall <key> → retrieves it. Omitting args starts interactive mode. Usage: /pantheon-remember"
-agent: "zeus"
----
-# /pantheon-remember — Memory Store & Recall
-
-**What:** Stores a memory entry via `memory_store()` or recalls via `memory_recall()`. Interactive mode if no arguments.
-
-## Usage
-
-```
-/pantheon-remember "Decisão: usar FastAPI em vez de Flask"         → Store (namespace: default)
-/pantheon-remember "Decisão: ..." --namespace=decisions             → Store em namespace específico
-/pantheon-remember recall "FastAPI decision"                        → Recall por key
-/pantheon-remember recall "FastAPI decision" --namespace=decisions   → Recall em namespace
-/pantheon-remember                                                   → Modo interativo (pergunta→store→pergunta recall)
-```
-
-## Behavior
-
-### Store mode (default)
-1. Chama `memory_store()` com o texto fornecido
-2. Metadata: preenche `agent`, `type`, `timestamp` automaticamente
-3. Pergunta: "Quer buscar algo relacionado? (y/n)"
-   - `y` → `memory_recall()` com extração automática de keywords
-   - `n` → fim
-
-### Recall mode
-1. Chama `memory_recall()` com a key fornecida
-2. Se não achar, tenta `memory_search()` como fallback
-
-### Interactive mode (no args)
-1. "O que você quer armazenar?"
-2. Recebe texto → `memory_store()`
-3. "Quer buscar algo relacionado?"
-4. Opcional: "Qual namespace?" (default: "default")
-
-## Output
-
-```
-✅ Stored — id: 42 | namespace: decisions | key: auto-generated
-📎 Related entries (memory_recall):
-  - #1: "Decisão: usar FastAPI..." (score: 0.92)
-```

package/commands/pantheon-search.md

@@ -1,48 +0,0 @@
----
-description: "Search memories by text with optional namespace filter, category, and result limit. /pantheon-search <query> --namespace=<ns> --top=<N> --stats. Usage: /pantheon-search"
-agent: "zeus"
----
-# /pantheon-search — Memory Search
-
-**What:** Searches memory entries via `memory_search()` with hybrid vector + keyword (RRF fusion). Supports namespace filter, result limit, and stats view.
-
-## Usage
-
-```
-/pantheon-search "FastAPI performance"                             → Busca em todos os namespaces
-/pantheon-search "FastAPI" --namespace=decisions                    → Busca só em decisions
-/pantheon-search "bug pattern" --namespace=sessions --top=10        → 10 resultados
-/pantheon-search --stats                                            → memory_stats() completo
-/pantheon-search --stats --namespace=decisions                       → Stats por namespace
-```
-
-## Flags
-
-| Flag | Default | Descrição |
-|------|---------|-----------|
-| `--namespace=<ns>` | `all` | Filtrar por namespace específico |
-| `--top=<N>` | `5` | Número de resultados (max 50) |
-| `--stats` | off | Mostrar estatísticas em vez de buscar |
-
-## Output (search)
-
-```
-🔍 Memory Search: "FastAPI performance"
-
-Results (namespace: all):
-  #1 │ decisions │ "Decisão: usar FastAPI..." │ score: 0.89 │ 2026-07-20
-  #2 │ repo      │ "FastAPI async patterns"  │ score: 0.74 │ 2026-07-18
-  #3 │ sessions  │ "FastAPI migration done"  │ score: 0.61 │ 2026-07-15
-
-Found 3 results in 2 namespaces (0.23s)
-```
-
-## Output (stats)
-
-```
-📊 Memory Stats
-Total entries: 142
-Namespaces: decisions (58), sessions (44), repo (28), default (12)
-Vector entries: 142
-DB size: 1.2 MB
-```

package/commands/pantheon-status.md

@@ -1,71 +0,0 @@
----
-description: "Show Pantheon version, model tier, and all available agents with status"
-agent: "zeus"
----
-# /pantheon-status — System Status & Agent Registry
-
-**What:** Displays the current Pantheon version, model tier (Pro/Free), and a full status table of all agents.
-**Usage:** `/pantheon-status`
-**Returns:** Version badge · model tier · agent table with availability
-**Use when:** You want to see what version of Pantheon is running and which agents are loaded
-
-## Output Format
-
-Produce EXACTLY this structure:
-
-```
-## 🏛️ Pantheon System Status
-
-**Version:** vX.Y.Z
-**Model Tier:** Pro (premium models) | Free (default models)
-**Platform:** <detect: opencode | claude | cursor | windsurf | vscode>
-**Agents loaded:** N of 14
-
-### Agent Registry
-
-| Agent | Role | Tier | Status |
-|-------|------|------|--------|
-| @zeus | Central orchestrator | default | ✅ |
-| @athena | Strategic planner | premium | ✅ |
-| @hermes | Backend (FastAPI) | default | ✅ |
-| @aphrodite | Frontend (React) | default | ✅ |
-| @demeter | Database | default | ✅ |
-| @themis | Quality & security | premium | ✅ |
-| @prometheus | Infrastructure | default | ✅ |
-| @apollo | Codebase discovery | fast | ✅ |
-| @iris | GitHub operations | fast | ✅ |
-| @mnemosyne | Memory bank | fast | ✅ |
-| @talos | Hotfixes | fast | ✅ |
-| @hephaestus | AI pipelines | default | ✅ |
-| @nyx | Observability | fast | ✅ |
-| @gaia | Remote sensing | fast | ✅ |
-
-### Commands Available
-
-`/pantheon` `/pantheon-status` `/audit` `/focus` `/deepwork` `/optimize` `/sketch` `/` `/` `/` `/pantheon-remember` `/pantheon-search` `/` `/pantheon-forget`
-
-### Health
-
-- ✅ All agents responding
-- ⚠️ N agents degraded (if any)
-- ❌ N agents unavailable (if any)
-```
-
-## How to Detect Version
-
-Read `package.json` → `"version"` field. If unavailable, read `plugin.json` → `"version"`.
-
-## How to Detect Model Tier
-
-- **Pro:** Any agent with `model_tier: premium` in `routing.yml` is using a premium model (athena, themis)
-- **Free:** All agents using `default` or `fast` tier models
-- Report based on what `routing.yml` declares for the active platform
-
-## How to Detect Platform
-
-Check for platform-specific markers:
-- `opencode.json` present → **OpenCode**
-- `.claude/` directory → **Claude Code**
-- `.cursor/` directory → **Cursor**
-- `.windsurf/` directory → **Windsurf**
-- `.vscode/` directory → **VS Code Copilot**

package/commands/pantheon-verify.md

@@ -1,17 +0,0 @@
----
-description: "Create verification plan before non-trivial changes"
-agent: zeus
----
-# /pantheon-verify — Verification Planning
-
-**What:** Plans an evidence path before non-trivial changes, defining how to verify correctness.
-**Usage:** `/pantheon-verify <change-description>`
-**Returns:** Verification plan with strategy, rollback, and acceptance criteria
-
-## When to Use
-- Before changes touching 3+ files
-- Before DB migrations or schema changes
-- Before API contract changes
-- Before dependency upgrades
-
-Delegates to @themis + @athena with `skill: verification-planning`.

package/scripts/check-version-consistency.sh

@@ -1,23 +0,0 @@
-#!/usr/bin/env bash
-# check-version-consistency.sh — Verifies all 3 manifests have the same version
-set -euo pipefail
-
-SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
-ROOT="$(dirname "$SCRIPT_DIR")"
-
-pkg_ver=$(node -p "require('$ROOT/package.json').version")
-plugin_ver=$(node -p "require('$ROOT/plugin.json').version")
-gh_plugin_ver=$(node -p "require('$ROOT/.github/plugin/plugin.json').version")
-
-echo "package.json: $pkg_ver"
-echo "plugin.json: $plugin_ver"
-echo ".github/plugin/plugin.json: $gh_plugin_ver"
-
-if [ "$pkg_ver" != "$plugin_ver" ] || [ "$pkg_ver" != "$gh_plugin_ver" ]; then
-  echo "❌ Version mismatch across manifests!"
-  echo ""
-  echo "Fix: run 'node scripts/versioning.mjs apply' to sync all manifests"
-  exit 1
-fi
-
-echo "✅ All versions match: $pkg_ver"

package/scripts/generate-routing-docs.mjs

@@ -1,104 +0,0 @@
-#!/usr/bin/env node
-/**
- * generate-routing-docs.mjs — Generate routing docs from routing.yml
- *
- * Reads the canonical routing.yml and outputs markdown documentation
- * to stdout. Pipe to a file to save: node generate-routing-docs.mjs > ROUTING.md
- */
-
-import { existsSync, readFileSync } from 'node:fs'
-import { dirname, join } from 'node:path'
-import { fileURLToPath } from 'node:url'
-import yaml from 'js-yaml'
-
-const __dirname = dirname(fileURLToPath(import.meta.url))
-const ROOT = join(__dirname, '..')
-
-const routingPath = join(ROOT, 'routing.yml')
-if (!existsSync(routingPath)) {
-  console.error('❌ routing.yml not found')
-  process.exit(1)
-}
-const routing = yaml.load(readFileSync(routingPath, 'utf8'))
-
-const agents = routing.agents || {}
-const matrix = routing.routing_matrix || []
-const handoffs = routing.handoffs || {}
-
-let doc = ''
-
-// Title
-doc += `# Pantheon Routing Reference\n\n`
-doc += `Auto-generated from \`routing.yml\` (v${routing.version || '?'})\n\n`
-doc += `${Object.keys(agents).length} agents · ${matrix.length} routing rules · ${Object.keys(handoffs).length} agent handoff groups\n\n`
-
-// Agent Registry
-doc += `---\n\n## Agent Registry\n\n`
-doc += `All ${Object.keys(agents).length} Pantheon agents with roles, model tiers, and capabilities.\n\n`
-doc += `| Agent | Role | Model Tier | User-Invocable | Delegates To | Skills |\n`
-doc += `|-------|------|-----------|----------------|-------------|--------|\n`
-
-for (const [name, info] of Object.entries(agents)) {
-  const role = info.role || '—'
-  const tier = info.model_tier || '—'
-  const invocable = info.user_invocable ? '✅ Yes' : '❌ No'
-  const delegates = (info.subagent_can_delegate_to || []).join(', ') || '—'
-  const agentSkills = (info.skills || []).map((s) => `\`${s}\``).join(', ') || '—'
-  doc += `| **@${name}** | ${role} | ${tier} | ${invocable} | ${delegates} | ${agentSkills} |\n`
-}
-
-// Routing Matrix
-doc += `\n---\n\n## Routing Matrix\n\n`
-doc += `Task category → primary agent mapping. Use this to decide which agent to invoke.\n\n`
-doc += `| Category | Primary Agent | Model Tier | Can Run Parallel With |\n`
-doc += `|----------|--------------|-----------|----------------------|\n`
-
-for (const entry of matrix) {
-  const primary = `@${entry.primary_agent}`
-  const tier = entry.model_tier || '—'
-  const parallel = entry.parallel_with ? entry.parallel_with.map((a) => `@${a}`).join(', ') : '—'
-  doc += `| ${entry.category} | ${primary} | ${tier} | ${parallel} |\n`
-}
-
-// Handoff Reference
-doc += `\n---\n\n## Handoff Reference\n\n`
-doc += `Structured delegation contracts between agents.\n\n`
-
-for (const [agentName, agentHandoffs] of Object.entries(handoffs)) {
-  doc += `### @${agentName}\n\n`
-  doc += `| Handoff | Target Agent | Model Tier | Description |\n`
-  doc += `|---------|-------------|-----------|-------------|\n`
-
-  for (const [key, handoff] of Object.entries(agentHandoffs)) {
-    if (typeof handoff !== 'object') continue
-    const target = `@${handoff.agent}`
-    const tier = handoff.model_tier || '—'
-    const desc = handoff.description || handoff.prompt?.substring(0, 80) || '—'
-    doc += `| **${key}** | ${target} | ${tier} | ${desc} |\n`
-  }
-  doc += `\n`
-}
-
-// Deployment platforms
-doc += `---\n\n## Deployment Platforms\n\n`
-doc += `Supported platforms and their configuration details.\n\n`
-doc += `| Platform | Directory | Format | Handoff Strategy |\n`
-doc += `|----------|-----------|--------|-----------------|\n`
-
-const PLATFORM_INFO = {
-  opencode: { dir: '.opencode/agents/', format: '.md with YAML', handoff: 'native (YAML)' },
-  claude: { dir: '.claude/agents/', format: '.md with YAML', handoff: 'embed (body)' },
-  cursor: { dir: '.cursor/rules/', format: '.mdc rule files', handoff: 'embed (body)' },
-  windsurf: { dir: '.windsurf/rules/', format: '.md rule files', handoff: 'embed (body)' },
-  cline: { dir: '.clinerules/', format: 'plain text', handoff: 'embed (body)' },
-  copilot: { dir: '.github/agents/', format: '.agent.md with YAML', handoff: 'native (YAML)' },
-  continue: { dir: '.continue/rules/', format: '.md rules', handoff: 'embed (body)' },
-}
-
-for (const [platform, info] of Object.entries(PLATFORM_INFO)) {
-  doc += `| **${platform}** | ${info.dir} | ${info.format} | ${info.handoff} |\n`
-}
-
-doc += `\n---\n\n*Generated on ${new Date().toISOString().split('T')[0]} from \`routing.yml\`*\n`
-
-console.log(doc)