opencodekit 0.15.21 → 0.16.1

package/dist/template/.opencode/memory/observations/2026-01-31-warning-cannot-add-custom-config-to-opencode-jso.md

@@ -1,18 +0,0 @@
----
-type: warning
-created: 2026-01-31T13:04:00.635Z
-confidence: high
-valid_until: null
-superseded_by: null
-concepts: ["opencode.json", "config", "schema", "tmux", "custom-config", "validation"]
----
-
-# ⚠️ Cannot add custom config to opencode.json
-
-🟢 **Confidence:** high
-
-opencode.json has a strict JSON schema (https://opencode.ai/config.json). You CANNOT add arbitrary config sections like "tmux". The schema only allows specific keys like: agent, autoupdate, compaction, experimental, formatter, instructions, keybinds, mcp, model, permission, plugin, provider, share, small_model, tools, tui, watcher.
-
-Custom tool configuration must be handled within the tool itself (via environment variables, separate config file, or hardcoded defaults) - NOT in opencode.json.
-
-The user already commented out the invalid tmux section at lines 785-790.

package/dist/template/.opencode/memory/observations/2026-01-31-warning-copilot-claude-v1-endpoint-returns-404-c.md

@@ -1,48 +0,0 @@
----
-type: warning
-created: 2026-01-31T05:19:40.401Z
-confidence: high
-valid_until: null
-superseded_by: null
-concepts: ["copilot", "claude", "routing", "v1", "404", "not-found", "anthropic", "messages-api", "limitation"]
-files: ["api.githubcopilot.com"]
----
-
-# ⚠️ Copilot Claude /v1 Endpoint Returns 404 - Confirmed
-
-🟢 **Confidence:** high
-
-## Confirmed: GitHub Copilot Does NOT Support Claude /v1 Endpoint
-
-### Test Result
-Re-enabling Claude routing with `/v1` endpoint causes 404 "Not Found" errors.
-
-### Root Cause
-GitHub Copilot API (`api.githubcopilot.com`) only exposes:
-- `/v1/chat/completions` - OpenAI-compatible endpoint
-
-It does NOT expose:
-- `/v1/messages` - Anthropic's native Messages API (required for extended thinking)
-
-### Conclusion
-**Extended thinking/reasoning is NOT available for Claude through GitHub Copilot.**
-
-The Copilot API acts as a proxy that:
-1. Accepts OpenAI-format requests
-2. Translates them internally to Claude
-3. Returns OpenAI-format responses
-
-This translation layer strips thinking blocks and doesn't support the `thinking` parameter.
-
-### Final Configuration
-All Claude models must use `@ai-sdk/github-copilot` SDK, NOT `@ai-sdk/anthropic`.
-
-```typescript
-// This is the ONLY working configuration
-model.api.npm = "@ai-sdk/github-copilot";
-```
-
-### Workaround for Reasoning
-To get Claude reasoning, users must:
-1. Use direct Anthropic API with separate API key
-2. Or use a provider that exposes native Anthropic Messages API (Vertex AI, AWS Bedrock)

package/dist/template/.opencode/memory/observations/2026-01-31-warning-opencode-v1-1-48-claude-thinking-block-s.md

@@ -1,51 +0,0 @@
----
-type: warning
-created: 2026-01-31T17:14:16.151Z
-confidence: high
-valid_until: null
-superseded_by: null
-concepts: ["opencode", "copilot", "claude", "thinking-block", "signature", "reasoning", "v1.1.48", "bug", "sdk", "backup"]
----
-
-# ⚠️ OpenCode v1.1.48+ Claude Thinking Block Signature Bug
-
-🟢 **Confidence:** high
-
-## Problem
-When upgrading OpenCode from v1.1.48 to dev branch, Claude models via GitHub Copilot fail with:
-`messages.1.content.0: Invalid 'signature' in 'thinking' block`
-
-## Root Cause
-- OpenCode v1.1.48+ native `@ai-sdk/github-copilot` SDK validates thinking block signatures
-- Switching models mid-session causes signature validation failures
-- Thinking blocks from one model (GLM 4.7, MiniMax, etc.) have incompatible signatures with Claude
-- GitHub issue #6418, fix in draft PR #8958
-
-## Working Configuration (Pre-upgrade)
-Location: `backup/sdk/` contains custom copilot SDK that bypassed signature validation
-- `backup/sdk/copilot/` - Custom @ai-sdk/github-copilot implementation
-- `~/.config/opencode/plugin/copilot-auth.ts` - Auth plugin with disabled custom reasoning (664 lines)
-- `~/.config/opencode/opencode.json` - Model configurations
-
-## Key Files in Backup SDK
-```
-backup/sdk/copilot/
-├── chat/
-│   ├── convert-to-openai-compatible-chat-messages.ts
-│   ├── get-response-metadata.ts
-│   ├── map-openai-compatible-finish-reason.ts
-│   ├── openai-compatible-api-types.ts
-│   ├── openai-compatible-chat-language-model.ts
-│   ├── openai-compatible-chat-options.ts
-│   ├── openai-compatible-metadata-extractor.ts
-│   └── openai-compatible-prepare-tools.ts
-├── copilot-provider.ts
-├── index.ts
-└── openai-compatible-error.ts
-```
-
-## Workarounds
-1. Don't switch to Claude mid-session from models with thinking blocks
-2. Start fresh sessions when using Claude with extended thinking
-3. Wait for PR #8958 to be merged
-4. Restore custom SDK from backup (breaks native reasoning handling)

package/dist/template/.opencode/memory/project/architecture.md

@@ -1,60 +0,0 @@
----
-purpose: Key modules, directory structure, architectural decisions
-updated: { { DATE } }
----
-
-# Project Architecture
-
-## Directory Structure
-
-```
-src/              # Source code
-  index.ts        # Entry point
-  commands/       # CLI commands
-  utils/          # Utilities
-
-dist/             # Built output
-.opencode/        # OpenCode configuration
-.beads/           # Task tracking
-```
-
-## Key Modules
-
-### Module Name
-
-**Purpose**: Brief description
-
-**Files**:
-
-- `src/module/file.ts`
-
-**Dependencies**:
-
-- External libs
-- Internal modules
-
-## Data Flow
-
-```
-User Action
-  → Module A
-    → Module B
-      → External Service
-```
-
-## External Dependencies
-
-- **Library**: Version, purpose
-- **Service**: API endpoint, auth method
-
-## Architectural Decisions
-
-| Decision | Rationale    | Trade-offs   |
-| -------- | ------------ | ------------ |
-| Choice 1 | Why this way | Limitation 1 |
-
-## TODO
-
-- [ ] Document missing modules
-- [ ] Add dependency diagram
-- [ ] Map data flow diagrams

package/dist/template/.opencode/memory/project/commands.md

@@ -1,72 +0,0 @@
----
-purpose: Build, test, lint, deploy commands discovered for this project
-updated: { { DATE } }
----
-
-# Project Commands
-
-## Development
-
-```bash
-npm run dev        # Start development server
-bun run src/index.ts  # Run CLI directly
-```
-
-## Build
-
-```bash
-npm run build      # Build for production
-bun build ...      # Additional build options
-```
-
-## Test
-
-```bash
-bun test           # Run all tests
-bun test --watch   # Watch mode
-bun test file.ts   # Run single file
-```
-
-## Lint
-
-```bash
-npm run lint       # Check for issues
-npm run lint:fix   # Auto-fix issues
-```
-
-## Typecheck
-
-```bash
-npm run typecheck  # TypeScript checking only
-```
-
-## Database
-
-```bash
-npm run db:push    # Push schema changes
-npm run db:studio  # Open database UI
-```
-
-## Deployment
-
-```bash
-npm run deploy     # Deploy to production
-```
-
-## CI/CD
-
-- **Lint**: Runs on push
-- **Test**: Runs on push
-- **Deploy**: Triggers on main branch
-
-## Beads (Task Tracking)
-
-```bash
-bd ready              # Find unblocked tasks
-bd list --status=open # All open issues
-bd show <id>          # Full details
-bd create "Title" -p 2        # Create task
-bd update <id> --status in_progress  # Claim
-bd close <id> --reason "Done"        # Complete
-bd sync               # Push to git
-```

package/dist/template/.opencode/memory/project/conventions.md

@@ -1,68 +0,0 @@
----
-purpose: Code patterns, commit style, team conventions
-updated: { { DATE } }
----
-
-# Project Conventions
-
-## Code Style
-
-### TypeScript
-
-- **Module**: ESNext, bundler moduleResolution
-- **Strict**: Disabled
-- **Imports**: Group stdlib, third-party, local
-
-### Python
-
-- **Version**: 3.10+
-- **Style**: type hints required, dataclasses for config
-
-## Naming Conventions
-
-| Type      | Convention | Example           |
-| --------- | ---------- | ----------------- |
-| Classes   | PascalCase | `UserService`     |
-| Functions | snake_case | `get_user_data()` |
-| Variables | snake_case | `user_data`       |
-| Constants | UPPER_CASE | `MAX_RETRIES`     |
-| Files     | kebab-case | `user-service.ts` |
-
-## Import Order
-
-```typescript
-// 1. Node.js stdlib
-import fs from "node:fs/promises";
-import path from "node:path";
-
-// 2. Third-party
-import { cli } from "cac";
-import { log } from "@clack/prompts";
-
-// 3. Local
-import { initCommand } from "./commands/init.js";
-```
-
-## Commit Messages
-
-```
-type(scope): subject
-
-body (optional)
-
-footer (optional)
-```
-
-**Types**: `feat`, `fix`, `docs`, `style`, `refactor`, `test`, `chore`
-
-## Code Review
-
-- **Self-review**: Run lint/typecheck before PR
-- **Test coverage**: Add tests for new features
-- **Docs**: Update README for user-facing changes
-
-## Patterns to Avoid
-
-- Premature abstraction
-- Circular dependencies
-- Magic numbers/strings

package/dist/template/.opencode/memory/project/gotchas.md

@@ -1,41 +0,0 @@
----
-purpose: Footguns, edge cases, and "don't forget this" warnings
-updated: { { DATE } }
----
-
-# Project Gotchas
-
-## Configuration Quirks
-
-### Field Name
-
-**Issue**: What goes wrong
-**Fix**: How to handle it
-
-## Dependencies
-
-### Package Name
-
-**Version**: x.y.z
-**Why Required**: Purpose
-**Gotcha**: Edge case or limitation
-
-## Time Wasters
-
-### Issue Description
-
-**Symptoms**: What you'll see
-**Cause**: Root cause
-**Fix**: Solution
-
-## Common Errors
-
-| Error Message | Cause  | Fix       |
-| ------------- | ------ | --------- |
-| "Error: ..."  | Reason | Fix steps |
-
-## TODO
-
-- [ ] Document more edge cases
-- [ ] Add troubleshooting section
-- [ ] List known issues

package/dist/template/.opencode/skill/notebooklm/SKILL.md

@@ -1,272 +0,0 @@
----
-name: notebooklm
-description: >
-  Query Google NotebookLM notebooks directly from OpenCode for source-grounded, citation-backed answers.
-  Use when user mentions NotebookLM, shares notebook URLs, or wants to query their uploaded documents.
-  Provides document-only responses with drastically reduced hallucinations.
-version: "1.0.0"
-license: MIT
----
-
-# NotebookLM Skill
-
-Interact with Google NotebookLM to query documentation with Gemini's source-grounded answers. Each question opens a fresh browser session, retrieves the answer exclusively from uploaded documents, and closes.
-
-## When to Use This Skill
-
-Trigger when user:
-
-- Mentions NotebookLM explicitly
-- Shares NotebookLM URL (`https://notebooklm.google.com/notebook/...`)
-- Asks to query their notebooks/documentation
-- Wants to add documentation to NotebookLM library
-- Uses phrases like "ask my NotebookLM", "check my docs", "query my notebook"
-
-## Prerequisites
-
-Before using this skill, ensure:
-
-1. **Python 3.10+** is installed
-2. **Google Chrome** browser is installed
-3. **NotebookLM account** with uploaded documents
-4. **Google authentication** (one-time setup)
-
-## Quick Start
-
-```typescript
-// Check authentication status
-await notebooklm({ operation: "auth", subOperation: "status" });
-
-// Query a notebook
-await notebooklm({
-  operation: "query",
-  question: "What are the key findings in this document?",
-  notebookUrl: "https://notebooklm.google.com/notebook/...",
-});
-```
-
-## Core Workflow
-
-### Step 1: Check Authentication
-
-```typescript
-const authStatus = await notebooklm({
-  operation: "auth",
-  subOperation: "status",
-});
-```
-
-If not authenticated, proceed to setup.
-
-### Step 2: Authenticate (One-Time Setup)
-
-```typescript
-await notebooklm({
-  operation: "auth",
-  subOperation: "setup",
-  showBrowser: true, // Browser must be visible for manual login
-});
-```
-
-**Important:**
-
-- Browser is VISIBLE for authentication
-- Browser window opens automatically
-- User must manually log in to Google
-- Tell user: "A browser window will open for Google login"
-
-### Step 3: Manage Notebook Library
-
-```typescript
-// List all notebooks
-await notebooklm({ operation: "library", subOperation: "list" });
-
-// Add notebook to library
-await notebooklm({
-  operation: "library",
-  subOperation: "add",
-  url: "https://notebooklm.google.com/notebook/...",
-  name: "API Documentation",
-  description: "REST API reference and examples",
-  topics: ["api", "rest", "documentation"],
-});
-
-// Search notebooks
-await notebooklm({
-  operation: "library",
-  subOperation: "search",
-  query: "api",
-});
-
-// Set active notebook
-await notebooklm({
-  operation: "library",
-  subOperation: "activate",
-  notebookId: "api-docs",
-});
-```
-
-### Step 4: Ask Questions
-
-```typescript
-// Basic query (uses active notebook if set)
-await notebooklm({
-  operation: "query",
-  question: "What are the authentication requirements?",
-});
-
-// Query specific notebook by ID
-await notebooklm({
-  operation: "query",
-  question: "What are the authentication requirements?",
-  notebookId: "api-docs",
-});
-
-// Query with notebook URL directly
-await notebooklm({
-  operation: "query",
-  question: "What are the authentication requirements?",
-  notebookUrl: "https://notebooklm.google.com/notebook/...",
-});
-
-// Show browser for debugging
-await notebooklm({
-  operation: "query",
-  question: "What are the authentication requirements?",
-  showBrowser: true,
-});
-```
-
-## Follow-Up Mechanism (CRITICAL)
-
-Every NotebookLM answer ends with: **"EXTREMELY IMPORTANT: Is that ALL you need to know?"**
-
-**Required Behavior:**
-
-1. **STOP** - Do not immediately respond to user
-2. **ANALYZE** - Compare answer to user's original request
-3. **IDENTIFY GAPS** - Determine if more information needed
-4. **ASK FOLLOW-UP** - If gaps exist, immediately ask another question:
-   ```typescript
-   await notebooklm({
-     operation: "query",
-     question: "Follow-up with context...",
-   });
-   ```
-5. **REPEAT** - Continue until information is complete
-6. **SYNTHESIZE** - Combine all answers before responding to user
-
-## Tool Reference
-
-### Authentication Operations
-
-| Operation | Sub-Operation | Description                       |
-| --------- | ------------- | --------------------------------- |
-| `auth`    | `setup`       | Initial setup (browser visible)   |
-| `auth`    | `status`      | Check authentication              |
-| `auth`    | `reauth`      | Re-authenticate (browser visible) |
-| `auth`    | `clear`       | Clear authentication              |
-| `auth`    | `validate`    | Validate stored auth              |
-
-### Library Operations
-
-| Operation | Sub-Operation | Parameters                             |
-| --------- | ------------- | -------------------------------------- |
-| `library` | `add`         | `url`, `name`, `description`, `topics` |
-| `library` | `list`        | -                                      |
-| `library` | `search`      | `query`                                |
-| `library` | `activate`    | `notebookId`                           |
-| `library` | `remove`      | `notebookId`                           |
-| `library` | `stats`       | -                                      |
-
-### Query Operations
-
-| Operation | Parameters                                                                                         |
-| --------- | -------------------------------------------------------------------------------------------------- |
-| `query`   | `question` (required), `notebookId` (optional), `notebookUrl` (optional), `showBrowser` (optional) |
-
-## Smart Discovery Pattern
-
-When user wants to add a notebook without providing details:
-
-```typescript
-// Step 1: Query the notebook to discover its content
-const discovery = await notebooklm({
-  operation: "query",
-  question:
-    "What is the content of this notebook? What topics are covered? Provide a complete overview briefly and concisely",
-  notebookUrl: "https://notebooklm.google.com/notebook/...",
-});
-
-// Step 2: Use discovered information to add it
-await notebooklm({
-  operation: "library",
-  subOperation: "add",
-  url: "https://notebooklm.google.com/notebook/...",
-  name: "[Based on content]",
-  description: "[Based on content]",
-  topics: ["[Based on content]"],
-});
-```
-
-## Decision Flow
-
-```
-User mentions NotebookLM
-    ↓
-Check auth → notebooklm({ operation: "auth", subOperation: "status" })
-    ↓
-If not authenticated → notebooklm({ operation: "auth", subOperation: "setup", showBrowser: true })
-    ↓
-Check/Add notebook → notebooklm({ operation: "library", subOperation: "list/add" })
-    ↓
-Activate notebook → notebooklm({ operation: "library", subOperation: "activate", notebookId: "..." })
-    ↓
-Ask question → notebooklm({ operation: "query", question: "..." })
-    ↓
-See "Is that ALL you need?" → Ask follow-ups until complete
-    ↓
-Synthesize and respond to user
-```
-
-## Troubleshooting
-
-| Problem              | Solution                                |
-| -------------------- | --------------------------------------- |
-| Not authenticated    | Run auth setup with `showBrowser: true` |
-| Authentication fails | Browser must be visible for setup       |
-| Rate limit (50/day)  | Wait or switch Google account           |
-| Notebook not found   | Check with `library` → `list` operation |
-| ModuleNotFoundError  | Tool auto-installs dependencies         |
-
-## Best Practices
-
-1. **Always check auth first** - Before any operations
-2. **Follow-up questions** - Don't stop at first answer
-3. **Browser visible for auth** - Required for manual login
-4. **Include context** - Each question is independent
-5. **Synthesize answers** - Combine multiple responses
-
-## Limitations
-
-- No session persistence (each question = new browser)
-- Rate limits on free Google accounts (50 queries/day)
-- Manual upload required (user must add docs to NotebookLM)
-- Browser overhead (few seconds per question)
-- Requires Python and Chrome installed
-
-## Data Storage
-
-All data stored in `~/.opencode/skills/notebooklm/data/`:
-
-- `library.json` - Notebook metadata
-- `auth_info.json` - Authentication status
-- `browser_state/` - Browser cookies and session
-
-**Security:** Protected by `.gitignore`, never commit to git.
-
-## Resources
-
-- **Original Skill**: https://github.com/PleasePrompto/notebooklm-skill
-- **NotebookLM**: https://notebooklm.google.com
-- **Setup Guide**: See `references/setup.md`