opencode-onboard 0.4.11 → 0.4.15

package/README.md

@@ -6,7 +6,7 @@
 
 **One command to prepare any codebase for AI agent workflows in OpenCode.**
 
-Works with [OpenCode](https://opencode.ai), [OpenCode Ensemble](https://github.com/hueyexe/opencode-ensemble), [OpenSpec](https://github.com/fission-ai/openspec), GitHub, Azure DevOps, or no tracker/PR platform at all.
+Works with [OpenCode](https://opencode.ai), [OpenCode Ensemble](https://github.com/hueyexe/opencode-ensemble), [OpenSpec](https://github.com/fission-ai/openspec), GitHub, Azure DevOps, or no tracker/PR platform at all.
 
 [![npm version](https://img.shields.io/npm/v/opencode-onboard?style=flat-square&color=black)](https://www.npmjs.com/package/opencode-onboard)
 [![npm downloads](https://img.shields.io/npm/dm/opencode-onboard?style=flat-square&color=black)](https://www.npmjs.com/package/opencode-onboard)
@@ -74,8 +74,8 @@ The CLI runs a 10-step onboarding wizard. It keeps the current step visible, plu
 | --------------------------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
 | **1. Source scope**               | Choose current repo or sibling source roots for code analysis                                                                                                        |
 | **2. Clean AI files**             | Detects existing `AGENTS.md`, `.cursorrules`, `CLAUDE.md`, `.agents/` etc. and removes them, preserves your `.agents/skills/`                                        |
-| **3. Choose platform**            | GitHub, Azure DevOps, or None                                                                                                                                       |
-| **4. Check platform CLI**         | Verifies `gh` (GitHub) or `az` + `azure-devops` (Azure DevOps), or skips checks when platform is None                                                              |
+| **3. Choose platform**            | GitHub, Azure DevOps, or None                                                                                                                                       |
+| **4. Check platform CLI**         | Verifies `gh` (GitHub) or `az` + `azure-devops` (Azure DevOps), or skips checks when platform is None                                                              |
 | **5. Copy scaffolding**           | Copies agents + built-in skills + bootstrap docs, writes source-roots metadata, applies AGENTS bootstrap patching, copies `skills-lock.json`, then runs `npx skills` |
 | **6. Init OpenSpec**              | Runs `npx @fission-ai/openspec init` silently for structured change management                                                                                       |
 | **7. Choose models**              | Fetches live model list from [models.dev](https://models.dev), lets you pick plan / build / fast models with cost indicators and canonical pricing                   |
@@ -100,7 +100,7 @@ Custom slash commands are installed into `.opencode/commands/` and are available
 | Command        | Description                                                                                           |
 | -------------- | ----------------------------------------------------------------------------------------------------- |
 | `/ob-init`        | Initialize the project. Asks greenfield vs brownfield, then activates the agent team. Supports skipping doc generation for new projects. |
-| `/ob-plan <url>`  | Parse a user story URL and produce a plan, proposal, specs, and tasks. Stops before implementation. Use platform mode, not `None`.   |
+| `/ob-plan <url>`  | Parse a user story URL and produce a plan, proposal, specs, and tasks. Stops before implementation. Use platform mode, not `None`.   |
 | `/ob-main <task>` | Quick direct implementation, no OpenSpec, no ensemble, no PRs. Just do it.                            |
 | `/ob-create-engineer <name> "<description>"` | Create a custom engineer agent from a description, with skills auto-installed from [skills.sh](https://www.skills.sh/) |
 | `/ob-create-architecture` | Generate or regenerate `ARCHITECTURE.md` from the codebase. Safe to rerun any time the architecture changes. |
@@ -123,14 +123,14 @@ devops-manager     lead/orchestrator, planning, PR lifecycle
 basic-engineer     implementation worker, ability-driven
 ```
 
-`basic-engineer` behavior is composed by abilities, not hardcoded role silos.
-Project-specific specialization comes from user-created custom engineers via `/ob-create-engineer`. During `/opsx-apply`, the lead should inspect the engineers that actually exist in `.agents/agents/`, prefer matching custom engineers, and fall back to `basic-engineer` only when no specialist is a clear fit.
+`basic-engineer` behavior is composed by abilities, not hardcoded role silos.
+Project-specific specialization comes from user-created custom engineers via `/ob-create-engineer`. During `/opsx-apply`, the lead should inspect the engineers that actually exist in `.opencode/agents/`, prefer matching custom engineers, and fall back to `basic-engineer` only when no specialist is a clear fit.
 
-### Skills, platform knowledge
-
-Skills define _what to know_. They provide project rules, platform behavior, and task-specific execution guidance. Agents auto-detect/load relevant skills; **you do not manually choose skills per prompt**.
-
-If you choose platform `None` during onboarding, no userstory or pull-request platform skills are injected into the workflow. The project works from direct conversation, local repo context, and optional OpenSpec artifacts only.
+### Skills, platform knowledge
+
+Skills define _what to know_. They provide project rules, platform behavior, and task-specific execution guidance. Agents auto-detect/load relevant skills; **you do not manually choose skills per prompt**.
+
+If you choose platform `None` during onboarding, no userstory or pull-request platform skills are injected into the workflow. The project works from direct conversation, local repo context, and optional OpenSpec artifacts only.
 
 Current loading model:
 
@@ -179,7 +179,7 @@ Models are fetched live from [models.dev](https://models.dev) (3000+ models, cac
 
 ## The pipeline
 
-When you give the lead agent a work item URL, execution follows this pipeline. If onboarding platform is `None`, skip the work item / PR stages and work directly from conversation plus optional OpenSpec artifacts:
+When you give the lead agent a work item URL, execution follows this pipeline. If onboarding platform is `None`, skip the work item / PR stages and work directly from conversation plus optional OpenSpec artifacts:
 
 ```
 devops-manager (load ob-global first)
@@ -191,7 +191,7 @@ devops-manager (load ob-global first)
                   ↓
              [confirm with user]
                   ↓
- basic-engineer + custom-engineer-* (parallel)
+ basic-engineer + *-engineer (parallel)
  claim tasks → load abilities → implement
                   ↓
        verify (tests/build/lint as needed)
@@ -218,7 +218,7 @@ Each agent runs in its own isolated git worktree via [OpenCode Ensemble](https:/
 
 ```
 your-project/
-├── AGENTS.md                        ← bootstrap mode, replaced after first "init"
+├── AGENTS.md                        ← bootstrap mode, replaced after first "/ob-init"
 ├── ARCHITECTURE.md                  ← prompt for agents to fill in from your codebase
 ├── DESIGN.md                        ← prompt for agents to fill in from your codebase
 ├── .opencode/

package/content/{.agents → .opencode}/agents/basic-engineer.md

@@ -1,8 +1,7 @@
 ---
 description: Basic Engineer Agent.
-mode: subagent
-color: #68A063
-temperature: 0.2
+mode: primary
+color: success
 permission:
   edit: allow
   bash: allow

package/content/{.agents → .opencode}/agents/devops-manager.md

@@ -25,10 +25,10 @@ Skills are located in `.agents/skills/`. Load required skills explicitly from co
 
 Always load `@ob-global` FIRST before any other skill.
 
-Platform skill selection must follow onboarding platform choice from CLI step:
-<!-- OB-PLATFORM-SKILLS-START -->
-- Platform-specific skill instructions are injected during onboarding copy step.
-<!-- OB-PLATFORM-SKILLS-END -->
+Platform skill selection must follow onboarding platform choice from CLI step:
+<!-- OB-PLATFORM-SKILLS-START -->
+- Platform-specific skill instructions are injected during onboarding copy step.
+<!-- OB-PLATFORM-SKILLS-END -->
 
 1. If the spawn prompt lists specific skills to load, read those `SKILL.md` files FIRST before any implementation
 2. Additionally, identify the platform from URLs or context
@@ -39,16 +39,16 @@ Examples of intent → skill mapping:
 - "PR has comments" or "review feedback" → load platform pullrequest observer skill
 - URL-based platform inference is fallback-only when onboarding metadata is unavailable
 
-Rules:
-- Platform selected in onboarding metadata takes precedence over URL inference when both exist
-- Never interact with a platform without loading the matching skill first
-- Follow skill instructions exactly, do not partially apply them
-- If no skill exists for the platform, report it as a blocker rather than improvising
-- Skills listed in the spawn prompt are MANDATORY, not optional
-
-<!-- OB-PLATFORM-MODE-START -->
-This project uses platform-integrated workflow modes described below.
-<!-- OB-PLATFORM-MODE-END -->
+Rules:
+- Platform selected in onboarding metadata takes precedence over URL inference when both exist
+- Never interact with a platform without loading the matching skill first
+- Follow skill instructions exactly, do not partially apply them
+- If no skill exists for the platform, report it as a blocker rather than improvising
+- Skills listed in the spawn prompt are MANDATORY, not optional
+
+<!-- OB-PLATFORM-MODE-START -->
+This project uses platform-integrated workflow modes described below.
+<!-- OB-PLATFORM-MODE-END -->
 
 ## Working Modes
 
@@ -65,11 +65,10 @@ This project uses platform-integrated workflow modes described below.
 3. Resolve platform from `.opencode/opencode-onboard.json` (`wizard.platform`) when available; fallback to URL inference only if missing/ambiguous
 4. Load the matching pullrequest skill for that resolved platform
 5. Capture screenshots of local running app if UI changes exist
-6. Read `.agents/session-log.json` if it exists, parse the JSON array and include a "Session Activity" section in the PR description with agent names, task counts, and skills used
-7. Commit and push the feature branch. If `## Source Roots` lists multiple roots, each root is a separate git repository, create and push the feature branch in EACH repo that has changes; never assume a single repo
-8. Create the PR following the skill instructions (one PR per repo that has changes)
-9. Post PR comment with screenshots and change summary
-10. Report PR URL to the lead
+6. Commit and push the feature branch. If `## Source Roots` lists multiple roots, each root is a separate git repository, create and push the feature branch in EACH repo that has changes; never assume a single repo
+7. Create the PR following the skill instructions (one PR per repo that has changes)
+8. Post PR comment with screenshots and change summary
+9. Report PR URL to the lead
 
 ### Feedback Mode (PR review loop)
 1. Load `@ob-global` first

package/content/.opencode/commands/ob-create-engineer.md

@@ -40,14 +40,13 @@ Example: `/ob-create-engineer frontend-engineer "A frontend engineer specialized
 
 4. **Create the agent file**
 
-   Create `.agents/agents/<name>.md` with this structure:
+    Create `.opencode/agents/<name>.md` with this structure:
 
    ```markdown
    ---
    description: <description>
-   mode: subagent
-   color: <pick a unique hex color>
-   temperature: 0.2
+   mode: primary
+   color: <pick a theme color: primary|secondary|accent|success|warning|error|info>
    permission:
      edit: allow
      bash: allow
@@ -65,17 +64,17 @@ Example: `/ob-create-engineer frontend-engineer "A frontend engineer specialized
    ## Workflow
 
    When spawned by the lead:
-   1. Call `team_tasks_list` immediately and identify your assigned task IDs.
-   2. Claim the first assigned task that is unblocked with `team_claim task_id:<id>`. If the first assigned task is blocked, claim the next assigned task whose dependencies are already `done`. Do not wait once you have an unblocked assigned task.
-   3. After claiming, load `@ob-global` first, then load mandatory ability `Guardrails`.
-   4. Load additional abilities from the `## Abilities` section as needed for the claimed task domain (for example: development, testing, infrastructure). Each ability can include one or more skills; load all relevant skills listed under each selected ability.
-   5. Send a short `team_message` to lead confirming the claimed task ID and loaded skills.
-   6. Implement the task following all loaded skill rules.
-   7. Call `team_tasks_complete task_id:<id>` after finishing that task.
-   8. Repeat until all currently assigned tasks are completed or blocked.
-   9. Message lead with results via `team_message`. Lead may assign more tasks, do NOT stop working or shut down until lead confirms no more tasks for you.
-   10. If lead sends new task IDs via `team_message`, treat them as new assignments and go back to step 1.
-   ```
+1. Call `team_tasks_list` immediately and identify your assigned task IDs.
+2. Claim the first assigned task that is unblocked with `team_claim task_id:<id>`. If the first assigned task is blocked, claim the next assigned task whose dependencies are already `done`. Do not wait once you have an unblocked assigned task.
+3. After claiming, load `@ob-global` first, then load mandatory ability `Guardrails`.
+4. Load additional abilities from the `## Abilities` section as needed for the claimed task domain (for example: development, testing, infrastructure). Each ability can include one or more skills; load all relevant skills listed under each selected ability.
+5. Send a short `team_message` to lead confirming the claimed task ID and loaded skills.
+6. Implement the task following all loaded skill rules.
+7. Call `team_tasks_complete task_id:<id>` after finishing that task.
+8. Repeat until all currently assigned tasks are completed or blocked.
+9. Message lead with results via `team_message`. Lead may assign more tasks, do NOT stop working or shut down until lead confirms no more tasks for you.
+10. If lead sends new task IDs via `team_message`, treat them as new assignments and go back to step 1.
+```
 
    Place the installed skills under the most relevant ability category:
    - **Development** — language frameworks, UI libraries, application code skills
@@ -88,13 +87,13 @@ Example: `/ob-create-engineer frontend-engineer "A frontend engineer specialized
 
    Add the new agent to the agents table in AGENTS.md:
    ```
-   | `<name>` | .agents/agents/<name>.md | <short role description> |
+    | `<name>` | .opencode/agents/<name>.md | <short role description> |
    ```
 
 6. **Show summary**
 
    Report:
-   - Agent file created at `.agents/agents/custom-engineer-<name>.md`
+    - Agent file created at `.opencode/agents/<name>-engineer.md`
    - Skills installed (list each with source)
    - How to use: "This agent will be spawned by the lead during `/opsx-apply` for tasks matching its specialty."
 

package/content/.opencode/commands/opsx-apply.md

@@ -117,15 +117,15 @@ Implement tasks from an OpenSpec change using the ensemble agent team.
        - Only shut down an agent when the board has no more tasks for its domain.
 
        Before spawning:
-       - scan `.agents/agents/` and list the engineers that actually exist in this project
+       - scan `.opencode/agents/` and list the engineers that actually exist in this project
        - exclude `devops-manager` from implementation selection
        - read each engineer's description and abilities
        - prefer the most specialized custom engineer whose description and abilities match the task
        - use `basic-engineer` only when no custom engineer is a clear fit or as a recovery fallback
-       - never spawn an engineer name that is not present in `.agents/agents/`
+       - never spawn an engineer name that is not present in `.opencode/agents/`
 
        REQUIRED assignment algorithm (do not skip):
-       1. Build candidate list from `.agents/agents/*.md` excluding `devops-manager`.
+       1. Build candidate list from `.opencode/agents/*.md` excluding `devops-manager`.
        2. Classify each task by domain using task text (api/backend, ui/frontend, infra/devops, testing/qa).
        3. For each task, score every candidate agent:
           - +3 if agent description explicitly matches domain
@@ -137,7 +137,7 @@ Implement tasks from an OpenSpec change using the ensemble agent team.
 
        HARD RULES:
        - NEVER assign a task to `basic-engineer` if a specialized agent has higher score.
-       - NEVER skip agent discovery from `.agents/agents/*.md`.
+       - NEVER skip agent discovery from `.opencode/agents/*.md`.
        - ALWAYS include assignment rationale only when it changes task routing or is needed to justify using `basic-engineer`.
 
        Skill loading is worker-driven:
@@ -217,7 +217,7 @@ Implement tasks from an OpenSpec change using the ensemble agent team.
           - `team_merge member:"<name>"`
           - If team_merge blocks on local changes: `git stash`, retry merge, `git stash pop`.
        6. **If ALL agents are shut down and tasks remain unassigned** (new domain, dependencies unblocked):
-          - Discover the remaining matching engineers from `.agents/agents/` and spawn a new wave (back to step 6d).
+          - Discover the remaining matching engineers from `.opencode/agents/` and spawn a new wave (back to step 6d).
        7. **If ALL tasks are done:** proceed to step 7.
 
        **ZERO PENDING TASKS GUARANTEE:** Before proceeding to step 7, call `team_tasks_list` and verify EVERY task is either `done` or `blocked`. If any task is `pending` and unassigned, assign it to an agent or spawn a new one. Never leave pending tasks orphaned.
@@ -264,7 +264,7 @@ Implement tasks from an OpenSpec change using the ensemble agent team.
 - ALWAYS call team_tasks_list after each agent reports done to check for remaining unassigned tasks
 - ALWAYS repeat the same literal task IDs in any task assignment message, never send a generic "claim your first task" without the actual IDs
 - NEVER send a start message that omits task IDs; if a task ID is missing from the start message, the agent cannot claim
-- ALWAYS discover engineers from `.agents/agents/` and prefer matching custom engineers over `basic-engineer`
+- ALWAYS discover engineers from `.opencode/agents/` and prefer matching custom engineers over `basic-engineer`
 - ALWAYS spawn agents sequentially (wait for each team_spawn result before the next), then send start messages to all of them together
 - ALWAYS set `claim_task` for the first unblocked task in each initial batch and instruct agents to claim before any other work
 - ALWAYS shut down + merge agents only when no more tasks remain for their domain

package/content/.opencode/skills/openspec-apply-change/SKILL.md

@@ -123,15 +123,15 @@ Implement tasks from an OpenSpec change using the ensemble agent team.
        - Only shut down an agent when the board has no more tasks for its domain.
 
        Before spawning:
-       - scan `.agents/agents/` and list the engineers that actually exist in this project
+       - scan `.opencode/agents/` and list the engineers that actually exist in this project
        - exclude `devops-manager` from implementation selection
        - read each engineer's description and abilities
        - prefer the most specialized custom engineer whose description and abilities match the task
        - use `basic-engineer` only when no custom engineer is a clear fit or as a recovery fallback
-       - never spawn an engineer name that is not present in `.agents/agents/`
+       - never spawn an engineer name that is not present in `.opencode/agents/`
 
        REQUIRED assignment algorithm (do not skip):
-       1. Build candidate list from `.agents/agents/*.md` excluding `devops-manager`.
+       1. Build candidate list from `.opencode/agents/*.md` excluding `devops-manager`.
        2. Classify each task by domain using task text (api/backend, ui/frontend, infra/devops, testing/qa).
        3. For each task, score every candidate agent:
           - +3 if agent description explicitly matches domain
@@ -143,7 +143,7 @@ Implement tasks from an OpenSpec change using the ensemble agent team.
 
        HARD RULES:
        - NEVER assign a task to `basic-engineer` if a specialized agent has higher score.
-       - NEVER skip agent discovery from `.agents/agents/*.md`.
+       - NEVER skip agent discovery from `.opencode/agents/*.md`.
        - ALWAYS include assignment rationale only when it changes task routing or is needed to justify using `basic-engineer`.
 
        Skill loading is worker-driven:
@@ -223,7 +223,7 @@ Implement tasks from an OpenSpec change using the ensemble agent team.
           - `team_merge member:"<name>"`
           - If team_merge blocks on local changes: `git stash`, retry merge, `git stash pop`.
        6. **If ALL agents are shut down and tasks remain unassigned** (new domain, dependencies unblocked):
-          - Discover the remaining matching engineers from `.agents/agents/` and spawn a new wave (back to step 6d).
+          - Discover the remaining matching engineers from `.opencode/agents/` and spawn a new wave (back to step 6d).
        7. **If ALL tasks are done:** proceed to step 7.
 
        **ZERO PENDING TASKS GUARANTEE:** Before proceeding to step 7, call `team_tasks_list` and verify EVERY task is either `done` or `blocked`. If any task is `pending` and unassigned, assign it to an agent or spawn a new one. Never leave pending tasks orphaned.
@@ -270,7 +270,7 @@ Implement tasks from an OpenSpec change using the ensemble agent team.
 - NEVER send a start message that omits task IDs; if a task ID is missing from the start message, the agent cannot claim
 - NEVER edit files between team_spawn and team_merge, team_merge blocks on overlapping local changes
 - ALWAYS add every task to the board before spawning, using multiple `team_tasks_add` calls when dependency wiring requires it
-- ALWAYS discover engineers from `.agents/agents/` and prefer matching custom engineers over `basic-engineer`
+- ALWAYS discover engineers from `.opencode/agents/` and prefer matching custom engineers over `basic-engineer`
 - ALWAYS spawn agents sequentially (wait for each team_spawn result before the next), then send start messages to all of them together
 - ALWAYS set `claim_task` for the first unblocked task in each initial batch and instruct agents to claim before any other work
 - ALWAYS shut down + merge agents only when no more tasks remain for their domain

package/content/AGENTS.md

@@ -6,7 +6,7 @@
 
 ## Trigger
 
-When the user says anything resembling initialization, "init", "initialize", "setup", "start", "bootstrap", "get started", "prepare", execute the steps below. Follow the greenfield/brownfield branching exactly.
+When the user says anything resembling initialization, "/ob-init", "initialize", "setup", "start", "bootstrap", "get started", "prepare", execute the steps below. Follow the greenfield/brownfield branching exactly.
 
 ---
 
@@ -174,10 +174,10 @@ Source scope is defined by mandatory `ob-global` skill.
 - Follow the generated `## Source Roots` section from that skill.
 - Do not duplicate source-scope rules here.
 
-## I Am the Lead, Full Workflow Ownership
-
-<!-- OB-PLATFORM-WORKFLOW-START -->
-When the user provides a work item URL or says "implement the plan" or "I've added comments to the PR", **I own the full lifecycle**. I load `ob-global` skill first, then the appropriate userstory skill, and use ensemble tools to coordinate the agent team.
+## I Am the Lead, Full Workflow Ownership
+
+<!-- OB-PLATFORM-WORKFLOW-START -->
+When the user provides a work item URL or says "implement the plan" or "I've added comments to the PR", **I own the full lifecycle**. I load `ob-global` skill first, then the appropriate userstory skill, and use ensemble tools to coordinate the agent team.
 
 Trigger patterns, I recognize ALL of these, exact wording does not matter:
 - User pastes or mentions a GitHub Issue URL → load `ob-userstory-gh` skill → parse issue → run `/opsx-propose` → confirm with user → run `/opsx-apply` → ship
@@ -186,20 +186,20 @@ Trigger patterns, I recognize ALL of these, exact wording does not matter:
 - `I've added comments to the PR` → read PR comments → fix → update PR
 - Any GitHub/Azure DevOps PR URL in a feedback/fix request (e.g. "check comments", "fix PR feedback") → run PR Feedback Loop
 
-**A GitHub or Azure DevOps URL anywhere in the user's message is always a trigger, regardless of surrounding words.**
-<!-- OB-PLATFORM-WORKFLOW-END -->
-
-**Never delegate without a plan. Default to specialists for implementation. If ensemble is clearly non-functional in the current session (idle teammate, no claim, or repeated spawn failure after one retry), stop forcing it: report the failure, then continue in the main session or ask the user whether to retry later.**
-
-## Engineer Selection
-
-Before spawning implementation workers:
-- Inspect `.agents/agents/*.md` and build the list of engineers that actually exist in this project.
-- Exclude `devops-manager` from implementation selection.
-- Prefer the most specialized custom engineer whose description and abilities clearly match the task domain.
-- Use `basic-engineer` only when no custom engineer is a clear fit or as a recovery fallback.
-- Never spawn engineer names that are not present in `.agents/agents/`.
-- When multiple engineers could fit, choose the narrower specialist before the generalist.
+**A GitHub or Azure DevOps URL anywhere in the user's message is always a trigger, regardless of surrounding words.**
+<!-- OB-PLATFORM-WORKFLOW-END -->
+
+**Never delegate without a plan. Default to specialists for implementation. If ensemble is clearly non-functional in the current session (idle teammate, no claim, or repeated spawn failure after one retry), stop forcing it: report the failure, then continue in the main session or ask the user whether to retry later.**
+
+## Engineer Selection
+
+Before spawning implementation workers:
+- Inspect `.opencode/agents/*.md` and build the list of engineers that actually exist in this project.
+- Exclude `devops-manager` from implementation selection.
+- Prefer the most specialized custom engineer whose description and abilities clearly match the task domain.
+- Use `basic-engineer` only when no custom engineer is a clear fit or as a recovery fallback.
+- Never spawn engineer names that are not present in `.opencode/agents/`.
+- When multiple engineers could fit, choose the narrower specialist before the generalist.
 
 ## Multi-Agent Execution, opencode-ensemble
 
@@ -218,10 +218,10 @@ Core tools used in this workflow:
 - **Max {{MAX_CONCURRENT_AGENTS}} truly concurrent agents.** All {{MAX_CONCURRENT_AGENTS}} must be spawned and running simultaneously, not sequentially. Spawn in waves if more than {{MAX_CONCURRENT_AGENTS}} are needed. Wait for wave N to finish before spawning wave N+1.
 - **Non-overlapping file domains.** Each agent owns exclusive directories. Two agents must NEVER touch the same file.
 - **Immediate shutdown on completion.** The moment an agent's domain has no more pending tasks → `team_shutdown` → `team_merge`. Keep agents alive if more tasks in their domain are pending (rolling batch).
-- **Rolling batch assignment.** Agents receive up to 3 tasks initially. When they complete a batch, lead assigns the next batch of up to 3 from the board. Never leave pending tasks orphaned.
-- **Stall detection at 5 minutes.** No commits after 5 min → nudge message → 2 min grace → force shutdown + respawn.
-- **Idle-without-claim is an earlier stall.** If a spawned teammate sits idle with no claimed task after a short wait, resend one short claim-only message with the exact task IDs. If still idle, force shutdown + respawn once with a shorter prompt. If the retry repeats the same failure, treat ensemble as unavailable for that session and stop recycling equivalent workers.
-- **Retry limit.** Max 3 retries per failing task → stop-and-report to user. Never retry indefinitely.
+- **Rolling batch assignment.** Agents receive up to 3 tasks initially. When they complete a batch, lead assigns the next batch of up to 3 from the board. Never leave pending tasks orphaned.
+- **Stall detection at 5 minutes.** No commits after 5 min → nudge message → 2 min grace → force shutdown + respawn.
+- **Idle-without-claim is an earlier stall.** If a spawned teammate sits idle with no claimed task after a short wait, resend one short claim-only message with the exact task IDs. If still idle, force shutdown + respawn once with a shorter prompt. If the retry repeats the same failure, treat ensemble as unavailable for that session and stop recycling equivalent workers.
+- **Retry limit.** Max 3 retries per failing task → stop-and-report to user. Never retry indefinitely.
 
 **Progress inspection commands (tell user explicitly after spawning):**
 - `team_status` for live team snapshot
@@ -236,19 +236,19 @@ If a teammate stalls due to model quota/rate-limit exhaustion:
 
 ---
 
-## Pipeline
-
-<!-- OB-PLATFORM-PIPELINE-START -->
-```
-devops-manager (lead mode)
-  → load ob-global + parse work item via skill
+## Pipeline
+
+<!-- OB-PLATFORM-PIPELINE-START -->
+```
+devops-manager (lead mode)
+  → load ob-global + parse work item via skill
         ↓
   openspec-propose
   → proposal.md + specs + tasks
         ↓
   [confirm with user]
         ↓
-basic-engineer + custom-engineer-* (parallel as needed)
+basic-engineer + *-engineer (parallel as needed)
   → claim tasks + load abilities + implement
         ↓
 devops-manager (ship mode)
@@ -269,11 +269,11 @@ devops-manager (ship mode)
 
 ```
 0. Run /quota to check remaining budget before spawning.
-1. Run /opsx-apply.
-   - Step 5b: classify cost tier, announce scope, ask user to confirm if ≥4 tasks.
-   - Lead adds all tasks to board.
-   - When dependencies exist, lead uses multiple `team_tasks_add` waves so later tasks can reference real task IDs returned by earlier waves.
-   - Lead discovers available engineers from `.agents/agents/*.md`, prefers matching custom engineers, then spawns engineers with initial batch of up to 3 tasks each (rolling batch model).
+1. Run /opsx-apply.
+   - Step 5b: classify cost tier, announce scope, ask user to confirm if ≥4 tasks.
+   - Lead adds all tasks to board.
+   - When dependencies exist, lead uses multiple `team_tasks_add` waves so later tasks can reference real task IDs returned by earlier waves.
+   - Lead discovers available engineers from `.opencode/agents/*.md`, prefers matching custom engineers, then spawns engineers with initial batch of up to 3 tasks each (rolling batch model).
    - Each engineer claims tasks, implements, completes, messages lead.
    - Lead assigns next batch (up to 3) to agents that report done. Repeat until board empty.
    - Lead merges each engineer branch after shutdown, then marks tasks done in tasks.md.
@@ -306,10 +306,10 @@ When user says "I've added comments to the PR" or asks to fix PR comments from P
 6. Wait for engineer results → team_shutdown + team_merge per engineer
 7. Run verification tasks (tests/build/lint) and fix blockers if any
 8. team_spawn devops-manager (ship mode) with "push + update PR threads" task ID + team_message "Start now"
-9. Wait → team_results → report what was updated
-10. team_cleanup
-```
-<!-- OB-PLATFORM-PIPELINE-END -->
+9. Wait → team_results → report what was updated
+10. team_cleanup
+```
+<!-- OB-PLATFORM-PIPELINE-END -->
 
 ---
 
@@ -319,10 +319,10 @@ All agents are universal, no project-specific knowledge. Platform and tech knowl
 
 | Agent | File | Role |
 |-------|------|------|
-| `devops-manager` | .agents/agents/devops-manager.md | Reads work items, creates PRs, handles review feedback |
-| `basic-engineer` | .agents/agents/basic-engineer.md | Generic implementation worker using ability-loaded skills |
+| `devops-manager` | .opencode/agents/devops-manager.md | Reads work items, creates PRs, handles review feedback |
+| `basic-engineer` | .opencode/agents/basic-engineer.md | Generic implementation worker using ability-loaded skills |
 
-User can add more custom engineer agents and run them in parallel. Keep behavior ability-driven via skill mappings. Custom engineers are the primary specialization mechanism; `basic-engineer` is the general fallback when no custom engineer is a clear fit.
+User can add more custom engineer agents and run them in parallel. Keep behavior ability-driven via skill mappings. Custom engineers are the primary specialization mechanism; `basic-engineer` is the general fallback when no custom engineer is a clear fit.
 
 Default `basic-engineer` abilities:
 
@@ -336,7 +336,7 @@ Default `basic-engineer` abilities:
 
 ## Skills
 
-Skills provide platform and tech-specific knowledge. Agents usually detect and load them automatically. Prefer auto-detection, but explicitly naming a skill in a spawn prompt is allowed when a workflow requires it or repeated misses show the agent is not loading the right context.
+Skills provide platform and tech-specific knowledge. Agents usually detect and load them automatically. Prefer auto-detection, but explicitly naming a skill in a spawn prompt is allowed when a workflow requires it or repeated misses show the agent is not loading the right context.
 
 `ob-global` is always loaded first, it provides baseline rules for all agents.
 
@@ -377,7 +377,7 @@ When `## Source Roots` lists multiple roots, each root is an independent git rep
 │   ├── agents/        # Agent definitions (universal, no project knowledge)
 │   │   ├── devops-manager.md
 │   │   ├── basic-engineer.md
-│   │   └── custom-engineer-*.md   # optional, user-defined workers
+│   │   └── *-engineer.md   # optional, user-defined workers
 │   └── skills/      # Skills (platform/tech specific knowledge)
 │       ├── ob-global/              ← baseline skill, load first
 │       ├── ob-default/            ← fallback skill

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "opencode-onboard",
-  "version": "0.4.11",
+  "version": "0.4.15",
   "description": "Prepare any brownfield codebase for AI agent workflows using OpenCode, OpenSpec, and ensemble orchestration.",
   "keywords": [
     "opencode",

package/src/commands/wizard.js

@@ -111,7 +111,7 @@ export async function runWizard(version) {
   console.log(chalk.bold.green('━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━'))
   console.log()
   console.log('  Open this project in OpenCode and type:')
-  console.log(chalk.bold('  "init"'))
+  console.log(chalk.bold('  "/ob-init"'))
   console.log()
   if (toGenerate.length > 0) {
     console.log(`  OpenCode will generate ${toGenerate.join(' and ')}`)

package/src/steps/copy/agents.js

@@ -123,7 +123,7 @@ function buildPipelineSection(platform) {
       '        ↓',
       '  [confirm with user when scope needs it]',
       '        ↓',
-      'basic-engineer + custom-engineer-* (parallel as needed)',
+      'basic-engineer + *-engineer (parallel as needed)',
       '  → claim tasks + load abilities + implement',
       '        ↓',
       'main session',
@@ -148,7 +148,7 @@ function buildPipelineSection(platform) {
       '   - Step 5b: classify cost tier, announce scope, ask user to confirm if ≥4 tasks.',
       '   - Lead adds all tasks to board.',
       '   - When dependencies exist, lead uses multiple `team_tasks_add` waves so later tasks can reference real task IDs returned by earlier waves.',
-      '   - Lead discovers available engineers from `.agents/agents/*.md`, prefers matching custom engineers, then spawns engineers with initial batch of up to 3 tasks each (rolling batch model).',
+       '   - Lead discovers available engineers from `.opencode/agents/*.md`, prefers matching custom engineers, then spawns engineers with initial batch of up to 3 tasks each (rolling batch model).',
       '   - Each engineer claims tasks, implements, completes, messages lead.',
       '   - Lead assigns next batch (up to 3) to agents that report done. Repeat until board empty.',
       '   - Lead merges each engineer branch after shutdown, then marks tasks done in tasks.md.',
@@ -177,7 +177,7 @@ function buildPipelineSection(platform) {
     '        ↓',
     '  [confirm with user]',
     '        ↓',
-    'basic-engineer + custom-engineer-* (parallel as needed)',
+    'basic-engineer + *-engineer (parallel as needed)',
     '  → claim tasks + load abilities + implement',
     '        ↓',
     'devops-manager (ship mode)',
@@ -202,7 +202,7 @@ function buildPipelineSection(platform) {
     '   - Step 5b: classify cost tier, announce scope, ask user to confirm if ≥4 tasks.',
     '   - Lead adds all tasks to board.',
     '   - When dependencies exist, lead uses multiple `team_tasks_add` waves so later tasks can reference real task IDs returned by earlier waves.',
-    '   - Lead discovers available engineers from `.agents/agents/*.md`, prefers matching custom engineers, then spawns engineers with initial batch of up to 3 tasks each (rolling batch model).',
+     '   - Lead discovers available engineers from `.opencode/agents/*.md`, prefers matching custom engineers, then spawns engineers with initial batch of up to 3 tasks each (rolling batch model).',
     '   - Each engineer claims tasks, implements, completes, messages lead.',
     '   - Lead assigns next batch (up to 3) to agents that report done. Repeat until board empty.',
     '   - Lead merges each engineer branch after shutdown, then marks tasks done in tasks.md.',
@@ -319,7 +319,7 @@ export async function patchAgentGuidance(platform, cwd = process.cwd()) {
 }
 
 export async function patchDevopsManagerMd(platform, cwd = process.cwd()) {
-  const devopsPath = path.join(cwd, '.agents', 'agents', 'devops-manager.md')
+  const devopsPath = path.join(cwd, '.opencode', 'agents', 'devops-manager.md')
   if (!await fse.pathExists(devopsPath)) return
 
   const resolved = platform === 'azure' ? 'azure' : platform === 'none' ? 'none' : 'github'

package/src/steps/copy/agents.test.js

@@ -30,8 +30,8 @@ describe('platform patching', () => {
   })
 
   it('patches devops-manager for none mode without platform skills', async () => {
-    const source = path.join(process.cwd(), 'content', '.agents', 'agents', 'devops-manager.md')
-    const dest = path.join(tmpDir, '.agents', 'agents', 'devops-manager.md')
+    const source = path.join(process.cwd(), 'content', '.opencode', 'agents', 'devops-manager.md')
+    const dest = path.join(tmpDir, '.opencode', 'agents', 'devops-manager.md')
     await fse.ensureDir(path.dirname(dest))
     await fse.copyFile(source, dest)
 

package/src/steps/models/index.js

@@ -12,7 +12,7 @@ export async function chooseModels() {
 
   if (!rawModels) {
     warn('Could not fetch models (offline and no cache). Skipping model selection.');
-    warn('Set models later in .agents/agents/<name>.md and .opencode/opencode.json');
+    warn('Set models later in .opencode/agents/<name>.md and .opencode/opencode.json');
     return;
   }
 
@@ -26,7 +26,7 @@ export async function chooseModels() {
   success(`${models.length} models available`);
   console.log();
   info('Cost indicators: [$] cheap  [$$] mid  [$$$] expensive');
-  info('Type to search. Change selections later in .agents/agents/ and .opencode/opencode.json');
+  info('Type to search. Change selections later in .opencode/agents/ and .opencode/opencode.json');
   console.log();
 
   for (const line of modelsPreset.roles.plan.info) info(line);
@@ -41,12 +41,12 @@ export async function chooseModels() {
   const fastModel = await pickModel(modelsPreset.roles.fast.prompt, models);
   console.log();
 
-  const agentsDir = path.join(process.cwd(), '.agents', 'agents');
+  const agentsDir = path.join(process.cwd(), '.opencode', 'agents');
   await writeModelsToConfigs({ planModel, buildModel, fastModel, agentsDir, preset: modelsPreset });
 
   console.log();
   warn('Make sure you have API access to the selected models.');
-  warn('Change them anytime in .agents/agents/<name>.md and .opencode/opencode.json');
+  warn('Change them anytime in .opencode/agents/<name>.md and .opencode/opencode.json');
 
   return { planModel, buildModel, fastModel };
 }

package/src/steps/openspec/ensemble.js

@@ -51,12 +51,12 @@ export const ENSEMBLE_SECTION = `6. **Implement via ensemble team**
       - Only shut down an agent when the board has no more tasks for its domain.
 
       Before spawning:
-      - scan \`.agents/agents/\` and list the engineers that actually exist in this project
+      - scan \`.opencode/agents/\` and list the engineers that actually exist in this project
       - exclude \`devops-manager\` from implementation selection
       - read each engineer's description and abilities
       - prefer the most specialized custom engineer whose description and abilities match the task
       - use \`basic-engineer\` only when no custom engineer is a clear fit or as a recovery fallback
-      - never spawn an engineer name that is not present in \`.agents/agents/\`
+      - never spawn an engineer name that is not present in \`.opencode/agents/\`
 
       Each \`team_spawn\` MUST include the agent field (required, causes NOT NULL error if omitted).
 
@@ -131,7 +131,7 @@ export const ENSEMBLE_SECTION = `6. **Implement via ensemble team**
           - \`team_merge member:"<name>"\`
           - If team_merge blocks on local changes: \`git stash\`, retry merge, \`git stash pop\`.
       6. **If ALL agents are shut down and tasks remain unassigned** (new domain, dependencies unblocked):
-          - Discover the remaining matching engineers from \`.agents/agents/\` and spawn a new wave (back to step 6d).
+          - Discover the remaining matching engineers from \`.opencode/agents/\` and spawn a new wave (back to step 6d).
       7. **If ALL tasks are done:** proceed to step 7.
       If a teammate reports rate-limit/quota/token exhaustion, immediately shutdown that teammate and respawn with an available model.
 
@@ -176,7 +176,7 @@ export const ENSEMBLE_SECTION = `6. **Implement via ensemble team**
 - NEVER send a start message that omits task IDs; if a task ID is missing from the start message, the agent cannot claim
 - NEVER edit files between team_spawn and team_merge, team_merge blocks on overlapping local changes
 - ALWAYS add every task to the board before spawning, using multiple \`team_tasks_add\` calls when dependency wiring requires it
-- ALWAYS discover engineers from \`.agents/agents/\` and prefer matching custom engineers over \`basic-engineer\`
+- ALWAYS discover engineers from \`.opencode/agents/\` and prefer matching custom engineers over \`basic-engineer\`
 - ALWAYS spawn agents sequentially (wait for each team_spawn result before the next), then send start messages to all of them together
 - ALWAYS set \`claim_task\` for the first unblocked task in each initial batch and instruct agents to claim before any other work
 - ALWAYS shut down + merge agents only when no more tasks remain for their domain

package/src/steps/openspec/ensemble.test.js

@@ -92,7 +92,7 @@ describe('ENSEMBLE_SECTION dependency guidance', () => {
   })
 
   it('prefers discovered custom engineers over hardcoded role inventory', () => {
-    expect(ENSEMBLE_SECTION).toContain('scan `.agents/agents/` and list the engineers that actually exist in this project')
+    expect(ENSEMBLE_SECTION).toContain('scan \`.opencode/agents/\` and list the engineers that actually exist in this project')
     expect(ENSEMBLE_SECTION).toContain('prefer the most specialized custom engineer')
     expect(ENSEMBLE_SECTION).not.toContain('front-engineer: UI, components, framework skills')
     expect(ENSEMBLE_SECTION).not.toContain('team_spawn name:"front" agent:"front-engineer"')

package/src/steps/openspec/index.js

@@ -1,6 +1,6 @@
 import { execa } from "execa"
 import path from "node:path"
-import { commandExists, error, header, info, loading, success, warn} from "../../utils/exec.js"
+import { commandExists, error, header, info, loading, success, warn } from "../../utils/exec.js"
 import { APPLY_TARGETS, patchApplyFile } from "./ensemble.js"
 
 export const openspecSteps = {
@@ -58,7 +58,7 @@ async function install() {
     success("OpenSpec installed")
   } catch (err) {
     error(`Failed to run openspec install: ${err.message}`)
-  }  
+  }
 }
 
 async function check() {
@@ -77,7 +77,7 @@ async function init() {
   try {
     const result = await execa(
       "npx",
-      ["@fission-ai/openspec", "init", "--tools", "opencode", "--force"],
+      ["@fission-ai/openspec", "/ob-init", "--tools", "opencode", "--force"],
       {
         cwd: process.cwd(),
         stdio: "pipe",

package/src/steps/optimization/codegraph.js

@@ -45,7 +45,7 @@ export async function fixCodegraphConfig() {
   if (rogueMcp) {
     for (const entry of Object.values(rogueMcp)) {
       if (Array.isArray(entry.command) && entry.command[0] === 'codegraph') {
-        entry.command = ['npx', ...entry.command]
+        entry.command = ['npx', '@colbymchenry/codegraph', ...entry.command.slice(1)]
       }
     }
     correctContent.mcp = { ...(correctContent.mcp || {}), ...rogueMcp }
@@ -69,7 +69,7 @@ export async function installCodegraph(options = {}) {
   try {
     const installResult = await execa(
       'npx',
-      ['@colbymchenry/codegraph', 'install', '--target=opencode', `--location=${location}`, '--yes'],
+      ['--yes', '@colbymchenry/codegraph', 'install', '--target=opencode', `--location=${location}`, '--yes'],
       {
         cwd: process.cwd(),
         reject: false,
@@ -98,7 +98,7 @@ export async function installCodegraph(options = {}) {
   loading('initializing codegraph project index...')
 
   try {
-    const initResult = await execa('npx', ['codegraph', 'init'], {
+    const initResult = await execa('npx', ['@colbymchenry/codegraph', 'init', '-i'], {
       cwd: process.cwd(),
       reject: false,
       stdio: 'pipe',

package/src/steps/optimization/codegraph.test.js

@@ -55,7 +55,7 @@ describe('fixCodegraphConfig()', () => {
     expect(result).toBe(true)
     expect(fs.existsSync(path.join(tmpDir, 'opencode.jsonc'))).toBe(false)
     const readResult = await fse.readJson(path.join(opencodeDir, 'opencode.json'))
-    expect(readResult.mcp.codegraph).toEqual({ command: ['npx', 'codegraph', 'serve', '--mcp'] })
+    expect(readResult.mcp.codegraph).toEqual({ command: ['npx', '@colbymchenry/codegraph', 'serve', '--mcp'] })
     expect(readResult.plugin).toEqual(["opencode-plugin-openspec@latest"])
   })
 
@@ -74,7 +74,7 @@ describe('fixCodegraphConfig()', () => {
     expect(result).toBe(true)
     expect(fs.existsSync(path.join(tmpDir, 'opencode.jsonc'))).toBe(false)
     const readResult = await fse.readJson(path.join(tmpDir, '.opencode', 'opencode.json'))
-    expect(readResult.mcp.codegraph.command).toEqual(['npx', 'codegraph', 'serve', '--mcp'])
+    expect(readResult.mcp.codegraph.command).toEqual(['npx', '@colbymchenry/codegraph', 'serve', '--mcp'])
   })
 
   it('handles JSONC with comments', async () => {
@@ -95,7 +95,7 @@ describe('fixCodegraphConfig()', () => {
     expect(result).toBe(true)
     expect(fs.existsSync(path.join(tmpDir, 'opencode.jsonc'))).toBe(false)
     const readResult = await fse.readJson(path.join(opencodeDir, 'opencode.json'))
-    expect(readResult.mcp.codegraph.command).toEqual(['npx', 'codegraph', 'serve', '--mcp'])
+    expect(readResult.mcp.codegraph.command).toEqual(['npx', '@colbymchenry/codegraph', 'serve', '--mcp'])
   })
 
   it('handles JSONC with URLs containing //', async () => {
@@ -118,7 +118,7 @@ describe('fixCodegraphConfig()', () => {
     expect(result).toBe(true)
     expect(fs.existsSync(path.join(tmpDir, 'opencode.jsonc'))).toBe(false)
     const readResult = await fse.readJson(path.join(opencodeDir, 'opencode.json'))
-    expect(readResult.mcp.codegraph.command).toEqual(['npx', 'codegraph', 'serve', '--mcp'])
+    expect(readResult.mcp.codegraph.command).toEqual(['npx', '@colbymchenry/codegraph', 'serve', '--mcp'])
   })
 
   it('removes unparseable opencode.jsonc, warns, and returns false', async () => {
@@ -143,7 +143,7 @@ describe('fixCodegraphConfig()', () => {
 
     expect(result).toBe(true)
     const readResult = await fse.readJson(path.join(tmpDir, '.opencode', 'opencode.json'))
-    expect(readResult.mcp.codegraph.command).toEqual(['npx', 'codegraph', 'serve', '--mcp'])
+    expect(readResult.mcp.codegraph.command).toEqual(['npx', '@colbymchenry/codegraph', 'serve', '--mcp'])
     expect(fs.existsSync(path.join(tmpDir, 'opencode.jsonc'))).toBe(false)
   })
 

package/content/.agents/session-log.json

@@ -1,41 +0,0 @@
-[
-  {
-    "ts": "2026-05-04T21:04:09.899Z",
-    "agent": "lead",
-    "member": null,
-    "agentType": null,
-    "team": null,
-    "action": "started",
-    "sessionId": "ses_20b31c36cffeBfsYI2hJ1BNcM5"
-  },
-  {
-    "ts": "2026-05-04T21:04:26.180Z",
-    "agent": "lead",
-    "member": null,
-    "agentType": null,
-    "team": null,
-    "action": "skill-loaded",
-    "skill": "ob-userstory-gh",
-    "source": "read-skill-file"
-  },
-  {
-    "ts": "2026-05-04T21:04:47.306Z",
-    "agent": "lead",
-    "member": null,
-    "agentType": null,
-    "team": null,
-    "action": "completed",
-    "filesEdited": 0,
-    "skills": [
-      "ob-userstory-gh"
-    ],
-    "usage": {
-      "inputTokensReported": null,
-      "outputTokensReported": null,
-      "totalTokensReported": null,
-      "tokenEstimateLow": 3469,
-      "tokenEstimateHigh": 6940,
-      "method": "heuristic"
-    }
-  }
-]

package/content/.opencode/plugins/session-log.js

@@ -1,523 +0,0 @@
-import fs from "node:fs"
-import path from "node:path"
-
-const LOG_FILE = ".agents/session-log.json"
-const SPAWN_MATCH_WINDOW_MS = 30000
-const UNMATCHED_SESSION_WARN_MS = 30000
-const DEBUG = process.env.SESSION_LOG_DEBUG === "true"
-
-// Per-session state
-const sessionState = new Map()
-
-// Lead session -> current team name
-const leadTeamBySession = new Map()
-
-// Pending spawn records waiting for a session.created match
-// Each entry: { leadSessionId, at, member, agentType, teamName, spawnedSessionId, intendedSkills }
-const pendingSpawns = []
-
-// spawnedSessionId -> pending spawn (direct correlation fast path)
-const pendingSpawnBySessionId = new Map()
-
-// Team -> completed session snapshots
-const completedByTeam = new Map()
-
-function ts() {
-  return new Date().toISOString()
-}
-
-function nowMs() {
-  return Date.now()
-}
-
-function appendEntry(directory, entry) {
-  const logPath = path.join(directory, LOG_FILE)
-  const dir = path.dirname(logPath)
-  if (!fs.existsSync(dir)) fs.mkdirSync(dir, { recursive: true })
-
-  let entries = []
-  if (fs.existsSync(logPath)) {
-    try { entries = JSON.parse(fs.readFileSync(logPath, "utf8")) } catch { /* ignore parse errors */ }
-  }
-  entries.push(entry)
-  fs.writeFileSync(logPath, JSON.stringify(entries, null, 2), "utf8")
-}
-
-function resolveAgentName(session) {
-  const agentPath = session?.agent
-  if (agentPath) {
-    const base = path.basename(agentPath, ".md")
-    if (base) return base
-  }
-  return "lead"
-}
-
-function resolveModel(session) {
-  // Try common field names used by OpenCode session objects
-  return session?.model || session?.modelId || session?.model_id || null
-}
-
-function addSkillToState(state, skillName) {
-  if (!skillName || !state) return false
-  if (!state.skills) state.skills = new Set()
-  if (state.skills.has(skillName)) return false
-  state.skills.add(skillName)
-  return true
-}
-
-// Extract skill names hinted in a spawn prompt by scanning for known skill-like tokens.
-// Matches strings that look like skill directory names (kebab-case words after "skill" keyword
-// or adjacent to known skill name patterns).
-function extractIntendedSkills(prompt) {
-  if (!prompt || typeof prompt !== "string") return []
-  const skills = new Set()
-
-  // Match explicit skill name mentions: e.g. "ob-pullrequest-gh", "browser-automation"
-  const kebabPattern = /\b([a-z][a-z0-9]*(?:-[a-z0-9]+){1,})\b/g
-  let m
-  while ((m = kebabPattern.exec(prompt)) !== null) {
-    const candidate = m[1]
-    // Filter to plausible skill names: 2+ segments, known prefixes or suffixes
-    if (
-      candidate.startsWith("ob-") ||
-      candidate.startsWith("openspec-") ||
-      candidate.startsWith("browser-") ||
-      candidate.endsWith("-gh") ||
-      candidate.endsWith("-az") ||
-      candidate.endsWith("-automation") ||
-      candidate.endsWith("-change") ||
-      candidate.endsWith("-engineer") ||
-      candidate.endsWith("-manager") ||
-      candidate.endsWith("-auditor")
-    ) {
-      skills.add(candidate)
-    }
-  }
-  return Array.from(skills).sort()
-}
-
-function toNum(v) {
-  if (typeof v === "number" && Number.isFinite(v)) return v
-  if (typeof v === "string" && v.trim() !== "" && !Number.isNaN(Number(v))) return Number(v)
-  return null
-}
-
-function extractReportedTokens(obj) {
-  const out = { input: 0, output: 0, total: 0, found: false }
-  const visited = new Set()
-
-  function walk(node) {
-    if (!node || typeof node !== "object") return
-    if (visited.has(node)) return
-    visited.add(node)
-
-    for (const [k, v] of Object.entries(node)) {
-      const key = String(k).toLowerCase()
-      const n = toNum(v)
-
-      if (n !== null) {
-        if ((key.includes("input") || key.includes("prompt")) && key.includes("token")) {
-          out.input += n
-          out.found = true
-        } else if ((key.includes("output") || key.includes("completion")) && key.includes("token")) {
-          out.output += n
-          out.found = true
-        } else if (key.includes("total") && key.includes("token")) {
-          out.total += n
-          out.found = true
-        }
-      }
-
-      if (v && typeof v === "object") walk(v)
-    }
-  }
-
-  walk(obj)
-  return out
-}
-
-function estimateTokens(state) {
-  const inTokens = Math.ceil((state.charIn || 0) / 4)
-  const outTokens = Math.ceil((state.charOut || 0) / 4)
-  const base = inTokens + outTokens + Math.max(0, (state.toolCalls || 0) * 20)
-  const low = Math.max(0, Math.floor(base * 0.7))
-  const high = Math.max(low, Math.ceil(base * 1.4))
-  return { low, high }
-}
-
-function usagePayload(state) {
-  const est = estimateTokens(state)
-  const reportedIn = state.reportedInputTokens || 0
-  const reportedOut = state.reportedOutputTokens || 0
-  const reportedTotalDirect = state.reportedTotalTokens || 0
-  const reportedTotalDerived = reportedIn + reportedOut
-  const reportedTotal = reportedTotalDirect || reportedTotalDerived || 0
-
-  let method = "heuristic"
-  if (reportedTotal > 0 && (est.low > 0 || est.high > 0)) method = "mixed"
-  else if (reportedTotal > 0) method = "reported"
-
-  return {
-    inputTokensReported: reportedIn || null,
-    outputTokensReported: reportedOut || null,
-    totalTokensReported: reportedTotal || null,
-    tokenEstimateLow: est.low,
-    tokenEstimateHigh: est.high,
-    method,
-  }
-}
-
-function buildCompletedSnapshot(state, sessionId) {
-  return {
-    sessionId,
-    agent: state.agentName,
-    member: state.member || null,
-    agentType: state.agentType || null,
-    team: state.teamName || null,
-    model: state.model || null,
-    skills: Array.from(state.skills || []).sort(),
-    intendedSkills: state.intendedSkills || [],
-    usage: usagePayload(state),
-    filesEdited: state.editCount || 0,
-  }
-}
-
-function buildTeamSkillsSummary(teamName) {
-  const rows = completedByTeam.get(teamName) || []
-  const byAgent = {}
-  for (const row of rows) {
-    const key = row.member || row.agent || "unknown"
-    if (!byAgent[key]) byAgent[key] = { agentType: row.agentType || null, model: row.model || null, skills: new Set(), intendedSkills: new Set() }
-    for (const s of row.skills || []) byAgent[key].skills.add(s)
-    for (const s of row.intendedSkills || []) byAgent[key].intendedSkills.add(s)
-  }
-  const out = {}
-  for (const [k, v] of Object.entries(byAgent)) {
-    const skills = Array.from(v.skills).sort()
-    const intendedSkills = Array.from(v.intendedSkills).sort()
-    const missingSkills = intendedSkills.filter(s => !v.skills.has(s))
-    out[k] = {
-      agentType: v.agentType,
-      model: v.model,
-      skills,
-      intendedSkills,
-      missingSkills: missingSkills.length > 0 ? missingSkills : undefined,
-    }
-  }
-  return out
-}
-
-function trackCompletedByTeam(snapshot) {
-  if (!snapshot.team) return
-  if (!completedByTeam.has(snapshot.team)) completedByTeam.set(snapshot.team, [])
-  completedByTeam.get(snapshot.team).push(snapshot)
-}
-
-function enqueuePendingSpawn(leadSessionId, args, spawnOutput) {
-  // Try to extract spawnedSessionId from team_spawn output for direct correlation
-  const spawnedSessionId =
-    spawnOutput?.sessionId ||
-    spawnOutput?.session_id ||
-    spawnOutput?.id ||
-    spawnOutput?.data?.sessionId ||
-    spawnOutput?.data?.session_id ||
-    spawnOutput?.data?.id ||
-    null
-
-  const record = {
-    leadSessionId,
-    at: nowMs(),
-    member: args?.name || null,
-    agentType: args?.agent || null,
-    teamName: leadTeamBySession.get(leadSessionId) || null,
-    spawnedSessionId,
-    intendedSkills: extractIntendedSkills(args?.prompt),
-  }
-
-  pendingSpawns.push(record)
-
-  if (spawnedSessionId) {
-    pendingSpawnBySessionId.set(spawnedSessionId, record)
-  }
-}
-
-function matchPendingSpawn(sessionId) {
-  const now = nowMs()
-
-  // Fast path: direct session ID correlation from team_spawn output
-  if (sessionId && pendingSpawnBySessionId.has(sessionId)) {
-    const record = pendingSpawnBySessionId.get(sessionId)
-    pendingSpawnBySessionId.delete(sessionId)
-    const idx = pendingSpawns.indexOf(record)
-    if (idx !== -1) pendingSpawns.splice(idx, 1)
-    return record
-  }
-
-  // Fallback: time-window heuristic (drop expired first)
-  for (let i = pendingSpawns.length - 1; i >= 0; i--) {
-    if (now - pendingSpawns[i].at > SPAWN_MATCH_WINDOW_MS) {
-      pendingSpawnBySessionId.delete(pendingSpawns[i].spawnedSessionId)
-      pendingSpawns.splice(i, 1)
-    }
-  }
-  if (pendingSpawns.length === 0) return null
-  return pendingSpawns.shift()
-}
-
-// Maps ensemble tool name -> function that extracts log entry fields from args
-const ENSEMBLE_TOOL_HANDLERS = {
-  team_create:         (args) => ({ action: "team-created", team: args.name }),
-  team_spawn:          (args) => ({ action: "teammate-spawned", name: args.name, agentType: args.agent }),
-  team_shutdown:       (args) => ({ action: "teammate-shutdown", name: args.name }),
-  team_merge:          (args) => ({ action: "teammate-merged", name: args.name }),
-  team_cleanup:        () => ({ action: "team-cleanup" }),
-  team_status:         () => ({ action: "team-status-checked" }),
-  team_results:        (args) => ({ action: "team-results-read", from: args.from }),
-  team_message:        (args) => ({ action: "team-message", to: args.to ?? "lead", preview: String(args.text ?? "").slice(0, 120) }),
-  team_broadcast:      (args) => ({ action: "team-broadcast", preview: String(args.text ?? "").slice(0, 120) }),
-  team_tasks_add:      (args) => ({ action: "tasks-added", count: Array.isArray(args.tasks) ? args.tasks.length : "?" }),
-  team_tasks_complete: (args) => ({ action: "task-completed", taskId: args.task_id }),
-  team_claim:          (args) => ({ action: "task-claimed", taskId: args.task_id }),
-}
-
-export const SessionLogPlugin = async ({ client, directory }) => {
-  return {
-    event: async ({ event }) => {
-      try {
-        if (event?.type === "session.created") {
-          const sessionId = event.properties?.id ?? event.properties?.sessionID
-          if (!sessionId) return
-
-          const res = await client.session.get({ path: { id: sessionId } })
-          const session = res?.data
-          const fallbackAgent = resolveAgentName(session)
-          const model = resolveModel(session)
-          const spawnMatch = matchPendingSpawn(sessionId)
-
-          const state = {
-            agentName: spawnMatch?.member || fallbackAgent,
-            member: spawnMatch?.member || null,
-            agentType: spawnMatch?.agentType || null,
-            teamName: spawnMatch?.teamName || null,
-            model,
-            intendedSkills: spawnMatch?.intendedSkills || [],
-            editCount: 0,
-            skills: new Set(),
-            startedAtMs: nowMs(),
-            toolCalls: 0,
-            charIn: 0,
-            charOut: 0,
-            reportedInputTokens: 0,
-            reportedOutputTokens: 0,
-            reportedTotalTokens: 0,
-            // Track whether this session was matched to a spawn record
-            spawnMatched: !!spawnMatch,
-          }
-
-          sessionState.set(sessionId, state)
-
-          const startedEntry = {
-            ts: ts(),
-            agent: state.agentName,
-            member: state.member,
-            agentType: state.agentType,
-            team: state.teamName,
-            model: state.model,
-            action: "started",
-            sessionId,
-          }
-          appendEntry(directory, startedEntry)
-
-          // Emit explicit teammate-registered entry when a spawn is matched
-          if (spawnMatch) {
-            appendEntry(directory, {
-              ts: ts(),
-              agent: state.agentName,
-              member: state.member,
-              agentType: state.agentType,
-              team: state.teamName,
-              model: state.model,
-              intendedSkills: state.intendedSkills,
-              action: "teammate-registered",
-              sessionId,
-              correlationMethod: spawnMatch.spawnedSessionId === sessionId ? "direct" : "time-window",
-            })
-          } else {
-            // No spawn match, schedule an unmatched-session warning
-            const capturedSessionId = sessionId
-            setTimeout(() => {
-              const s = sessionState.get(capturedSessionId)
-              if (!s || s.spawnMatched) return
-              appendEntry(directory, {
-                ts: ts(),
-                agent: s.agentName,
-                member: s.member,
-                team: s.teamName,
-                model: s.model,
-                action: "unmatched-session",
-                sessionId: capturedSessionId,
-                warning: "Session started with no matching team_spawn record. Agent identity may be inaccurate.",
-              })
-            }, UNMATCHED_SESSION_WARN_MS)
-          }
-        }
-
-        if (event?.type === "file.edited") {
-          const sessionId = event.properties?.sessionID ?? event.properties?.id
-          if (!sessionId) return
-          const state = sessionState.get(sessionId)
-          if (state) state.editCount++
-        }
-
-        if (event?.type === "session.idle") {
-          const sessionId = event.properties?.id ?? event.properties?.sessionID
-          if (!sessionId) return
-          const state = sessionState.get(sessionId)
-          if (!state) return
-
-          const skills = Array.from(state.skills || []).sort()
-          const usage = usagePayload(state)
-          appendEntry(directory, {
-            ts: ts(),
-            agent: state.agentName,
-            member: state.member,
-            agentType: state.agentType,
-            team: state.teamName,
-            model: state.model,
-            action: "completed",
-            filesEdited: state.editCount,
-            skills,
-            intendedSkills: state.intendedSkills,
-            usage,
-          })
-
-          trackCompletedByTeam(buildCompletedSnapshot(state, sessionId))
-          sessionState.delete(sessionId)
-        }
-      } catch  {
-        // ignore
-      }
-    },
-
-    "tool.execute.after": async (input, output) => {
-      try {
-        const sessionId = input?.sessionID ?? input?.session_id
-        if (!sessionId) return
-
-        const state = sessionState.get(sessionId)
-        if (!state) return
-
-        const tool = input?.tool
-        const args = input?.args ?? {}
-
-        state.toolCalls++
-        state.charIn += JSON.stringify(args).length
-        state.charOut += JSON.stringify(output ?? {}).length
-
-        const reportedIn = extractReportedTokens(input)
-        const reportedOut = extractReportedTokens(output)
-        if (reportedIn.found) {
-          state.reportedInputTokens += reportedIn.input
-          state.reportedOutputTokens += reportedIn.output
-          state.reportedTotalTokens += reportedIn.total
-        }
-        if (reportedOut.found) {
-          state.reportedInputTokens += reportedOut.input
-          state.reportedOutputTokens += reportedOut.output
-          state.reportedTotalTokens += reportedOut.total
-        }
-
-        if (DEBUG && !reportedIn.found && !reportedOut.found && tool !== "read") {
-          appendEntry(directory, {
-            ts: ts(),
-            agent: state.agentName,
-            action: "debug-no-token-metrics",
-            tool,
-          })
-        }
-
-        // Track skill loads via skill tool (primary)
-        if (tool === "skill") {
-          const skillName = input?.args?.name
-          const added = addSkillToState(state, skillName)
-          if (added) {
-            appendEntry(directory, {
-              ts: ts(),
-              agent: state.agentName,
-              member: state.member,
-              agentType: state.agentType,
-              team: state.teamName,
-              model: state.model,
-              action: "skill-loaded",
-              skill: skillName,
-              source: "skill-tool",
-            })
-          }
-          return
-        }
-
-        // Track skill loads via reading SKILL.md (fallback)
-        if (tool === "read") {
-          const filePath = input?.args?.filePath ?? ""
-          const match = filePath.match(/[/\\]skills[/\\]([^/\\]+)[/\\]SKILL\.md$/i)
-          if (match) {
-            const skillName = match[1]
-            const added = addSkillToState(state, skillName)
-            if (added) {
-              appendEntry(directory, {
-                ts: ts(),
-                agent: state.agentName,
-                member: state.member,
-                agentType: state.agentType,
-                team: state.teamName,
-                model: state.model,
-                action: "skill-loaded",
-                skill: skillName,
-                source: "read-skill-file",
-              })
-            }
-          }
-          return
-        }
-
-        // Track ensemble tool calls
-        const ensembleHandler = ENSEMBLE_TOOL_HANDLERS[tool]
-        if (!ensembleHandler) return
-
-        const entry = {
-          ts: ts(),
-          agent: state.agentName,
-          member: state.member,
-          agentType: state.agentType,
-          team: state.teamName,
-          model: state.model,
-          ...ensembleHandler(args),
-        }
-        appendEntry(directory, entry)
-
-        if (tool === "team_create") {
-          leadTeamBySession.set(sessionId, args?.name || null)
-          state.teamName = args?.name || state.teamName
-        }
-
-        if (tool === "team_spawn") {
-          enqueuePendingSpawn(sessionId, args, output)
-        }
-
-        if (tool === "team_cleanup") {
-          const teamName = state.teamName || leadTeamBySession.get(sessionId)
-          appendEntry(directory, {
-            ts: ts(),
-            agent: state.agentName,
-            model: state.model,
-            action: "team-skills-summary",
-            team: teamName || null,
-            byAgent: teamName ? buildTeamSkillsSummary(teamName) : {},
-          })
-        }
-      } catch { 
-        // ignore
-      }
-    }
-  }
-}