opencodekit 0.18.0 → 0.18.2

package/dist/index.js

@@ -18,7 +18,7 @@ var __require = /* @__PURE__ */ createRequire(import.meta.url);
 
 //#endregion
 //#region package.json
-var version = "0.18.0";
+var version = "0.18.2";
 
 //#endregion
 //#region src/utils/errors.ts

package/dist/template/.opencode/AGENTS.md

@@ -220,3 +220,60 @@ Files over ~500 lines become hard to maintain and review. Extract helpers, split
 - Cite concrete file paths and line numbers for non-trivial claims
 
 _Complexity is the enemy. Minimize moving parts._
+
+---
+
+## Memory System
+
+4-tier automated knowledge pipeline backed by SQLite + FTS5 (porter stemming).
+
+**Pipeline:** messages → capture → distillations (TF-IDF) → observations (curator) → LTM injection (system.transform)
+
+### Memory Tools
+
+```bash
+# Search observations (FTS5)
+memory-search({ query: "auth" })
+
+# Get full observation details
+memory-get({ ids: "42,45" })
+
+# Create observation
+observation({ type: "decision", title: "Use JWT", narrative: "..." })
+
+# Update memory file
+memory-update({ file: "research/findings", content: "..." })
+
+# Read memory file
+memory-read({ file: "research/findings" })
+
+# Admin operations
+memory-admin({ operation: "status" })
+memory-admin({ operation: "capture-stats" })
+memory-admin({ operation: "distill-now" })
+memory-admin({ operation: "curate-now" })
+```
+
+### Session Tools
+
+```bash
+# Search sessions by keyword
+find_sessions({ query: "auth", limit: 5 })
+
+# Read session messages
+read_session({ session_id: "abc123" })
+read_session({ session_id: "abc123", focus: "auth" })
+```
+
+### Directory Structure
+
+```
+.opencode/memory/
+├── project/           # Tacit knowledge (auto-injected)
+│   ├── user.md        # User preferences
+│   ├── tech-stack.md  # Framework, constraints
+│   └── gotchas.md     # Footguns, warnings
+├── research/          # Research notes
+├── handoffs/          # Session handoffs
+└── _templates/        # Document templates
+```

package/dist/template/.opencode/agent/scout.md

@@ -47,43 +47,6 @@ Find trustworthy external references quickly and return concise, cited guidance.
 - Cite every non-trivial claim
 - Prefer high-signal synthesis over long dumps
 
-## Source Quality Hierarchy
-
-| Rank | Source Type                                           | Tiebreaker                                     |
-| ---- | ----------------------------------------------------- | ---------------------------------------------- |
-| 1    | Official docs/specifications/release notes            | Use unless clearly outdated                    |
-| 2    | Library/framework source code and maintained examples | Prefer recent commits                          |
-| 3    | Maintainer-authored technical articles                | Check date, prefer <1 year                     |
-| 4    | Community blogs/posts                                 | Use only when higher-ranked sources are absent |
-
-## Output
-
-- Summary (2-5 bullets)
-- Recommended approach
-- Sources
-- Risks/tradeoffs
-
-# Scout Agent
-
-**Purpose**: Knowledge seeker — you find the signal in the noise of external information.
-
-> _"Good research doesn't dump facts; it creates actionable clarity."_
-
-## Identity
-
-You are a read-only research agent. You output concise recommendations backed by verifiable sources only.
-
-## Task
-
-Find trustworthy external references quickly and return concise, cited guidance.
-
-## Rules
-
-- Never modify project files
-- Never invent URLs; only use verified links
-- Cite every non-trivial claim
-- Prefer high-signal synthesis over long dumps
-
 ## Before You Scout
 
 - **Verify memory first**: Always check memory-search before external research

package/dist/template/.opencode/command/resume.md

@@ -49,7 +49,7 @@ If a handoff exists, read the latest one. It tells you:
 Also search previous sessions:
 
 ```typescript
-search_session({ query: "$ARGUMENTS" });
+find_sessions({ query: "$ARGUMENTS" });
 ```
 
 ## Phase 4: Load Artifacts

package/dist/template/.opencode/command/status.md

@@ -40,12 +40,11 @@ skill({ name: "beads" });
 
 ## Available Tools
 
-| Tool            | Use When                          |
-| --------------- | --------------------------------- |
-| `br`            | Task status and stats             |
-| `git`           | Git state and history             |
-| `list_sessions` | Recent sessions                   |
-| `action-queue`  | Pending approvals and ready tasks |
+| Tool            | Use When              |
+| --------------- | --------------------- |
+| `br`            | Task status and stats |
+| `git`           | Git state and history |
+| `find_sessions` | Recent sessions       |
 
 ## Phase 1: Gather State (Parallel)
 
@@ -64,8 +63,7 @@ git log --oneline -5
 ```
 
 ```typescript
-list_sessions({ since: "today", limit: 5 });
-action - queue({ op: "status" });
+find_sessions({ query: "today", limit: 5 });
 ```
 
 ---
@@ -89,12 +87,7 @@ GIT
   Recent:  [from git log]
 
 SESSIONS TODAY
-  [from list_sessions]
-
-ACTION QUEUE
-  Pending approvals: [from action-queue pending_approvals]
-  Ready tasks:       [from action-queue ready_tasks]
-  Idle workers:      [from action-queue idle_workers]
+  [from find_sessions]
 ```
 
 ---

package/dist/template/.opencode/plugin/README.md

@@ -6,36 +6,48 @@ Plugins in this directory extend OpenCode with project-specific behavior and too
 
 ```text
 plugin/
-├── memory.ts           # Memory DB maintenance + observation toasts
-├── sessions.ts         # Session tools (list/read/search/summarize)
-├── compaction.ts       # Compaction-time context recovery injection
-├── swarm-enforcer.ts   # Beads workflow enforcement and reminders
-├── skill-mcp.ts        # Skill-scoped MCP bridge (skill_mcp tools)
+├── memory.ts           # 4-tier automated memory system (capture → distill → curate → inject)
+├── sessions.ts         # Session search tools (find/read)
 ├── copilot-auth.ts     # GitHub Copilot provider/auth integration
-├── lib/
-│   ├── memory-db.ts    # SQLite + FTS5 memory backend
-│   └── notify.ts       # Shared notification helpers
-└── sdk/                # Copilot SDK adaptation code
+├── skill-mcp.ts        # Skill-scoped MCP bridge (skill_mcp tools)
+└── lib/
+    ├── memory-tools.ts       # 6 core memory tools (observation, search, get, read, update, timeline)
+    ├── memory-admin-tools.ts # Admin tool (memory-admin: 9 operations)
+    ├── memory-hooks.ts       # All hooks (event, idle, transforms, compaction)
+    ├── memory-helpers.ts     # Constants, compaction utilities, formatting
+    ├── memory-db.ts          # Barrel re-export for db/ modules
+    ├── capture.ts            # message.part.updated → temporal_messages
+    ├── distill.ts            # TF-IDF extraction, key sentence selection
+    ├── curator.ts            # Pattern-based knowledge extraction
+    ├── context.ts            # Token budget enforcement via messages.transform
+    ├── inject.ts             # Relevance-scored LTM injection via system.transform
+    ├── notify.ts             # Cross-platform notification helpers
+    └── db/
+        ├── types.ts          # All types + MEMORY_CONFIG
+        ├── schema.ts         # SQL schema, migrations, DB singleton
+        ├── observations.ts   # Observation CRUD + FTS5 search
+        ├── pipeline.ts       # Temporal messages + distillations + relevance scoring
+        └── maintenance.ts    # Memory files, FTS5, archiving, vacuum
 ```
 
 ## Plugin Responsibilities
 
 - `memory.ts`
-  - Optimizes FTS5 index on idle sessions
-  - Checkpoints WAL when needed
-  - Shows toast feedback for observation saves and session errors
+  - 4-tier automated knowledge system: temporal_messages → distillations → observations → memory_files
+  - Captures messages automatically via `message.part.updated` events
+  - Distills sessions on idle (TF-IDF, key sentence extraction)
+  - Curates observations from distillations via pattern matching
+  - Injects relevant knowledge into system prompt (BM25 _ recency _ confidence scoring)
+  - Manages context window via messages.transform (token budget enforcement)
+  - Merges compaction logic (beads, handoffs, project memory, knowledge)
+  - Provides 7 tools: observation, memory-search, memory-get, memory-read, memory-update, memory-timeline, memory-admin
 
 - `sessions.ts`
-  - Provides custom tools: `list_sessions`, `read_session`, `search_session`, `summarize_session`
-
-- `compaction.ts`
-  - Injects session continuity context during compaction
-  - Pulls memory/project/handoff context and recovery instructions
-
-- `swarm-enforcer.ts`
-  - Injects bead state and stage labels into system context
-  - Warns when implementation starts without a properly started bead
-  - Reminds to close/sync in-progress work on session idle
+  - Provides tools: `find_sessions`, `read_session`
+  - Direct SQLite access to OpenCode's session DB
+  - Multi-word AND search with relevance ranking
+  - 180-day time-bounded search
+  - Agentic `nextStep` guidance in results
 
 - `skill-mcp.ts`
   - Loads MCP configs from skills
@@ -48,9 +60,9 @@ plugin/
 
 ## Notes
 
-- `notification.ts.bak` is a backup file and not part of the active plugin set.
-- Keep plugin documentation aligned with actual files in this directory.
-- Prefer shared helpers in `lib/` over duplicated utilities across plugins.
+- OpenCode auto-discovers every `.ts` file in `plugin/` as a plugin — keep helper modules in `lib/`
+- Keep plugin documentation aligned with actual files in this directory
+- Prefer shared helpers in `lib/` over duplicated utilities across plugins
 
 ## References
 

package/dist/template/.opencode/skill/accessibility-audit/SKILL.md

@@ -1,12 +1,13 @@
 ---
 name: accessibility-audit
 description: Use when auditing UI components or pages for accessibility compliance, checking WCAG conformance, identifying keyboard navigation issues, color contrast problems, and pre-launch accessibility verification
+version: 1.0.0
+tags: [ui, code-quality]
+dependencies: []
 ---
 
 # Accessibility Audit Skill
 
-Audit UI accessibility against WCAG 2.1/2.2 guidelines.
-
 ## When to Use
 
 - Reviewing UI components for accessibility compliance
@@ -15,6 +16,11 @@ Audit UI accessibility against WCAG 2.1/2.2 guidelines.
 - Checking color contrast
 - Pre-launch accessibility verification
 
+## When NOT to Use
+
+- Non-UI work where accessibility isn't applicable.
+
+
 ## Core Workflow
 
 ### Phase 1: Visual Accessibility Analysis
@@ -179,7 +185,7 @@ Save audit reports to `.opencode/memory/design/accessibility/`
 
 ## Related Skills
 
-| Need           | Skill                 |
-| -------------- | --------------------- |
+| Need           | Skill             |
+| -------------- | ----------------- |
 | Design quality | `frontend-design` |
-| UI research    | `ui-ux-research`      |
+| UI research    | `ui-ux-research`  |

package/dist/template/.opencode/skill/agent-browser/SKILL.md

@@ -1,12 +1,21 @@
 ---
 name: agent-browser
 description: Browser automation CLI for AI agents using Playwright
-version: 1
+version: 1.0.0
+tags: [automation, mcp]
+dependencies: []
 ---
 
 # agent-browser
 
-Browser automation CLI for AI agents. Use when tasks require web browsing, scraping, form filling, or web interaction. Built on Playwright with Rust CLI + Node.js daemon architecture.
+## When to Use
+
+- When you need Playwright-based browser automation via the agent-browser CLI.
+
+## When NOT to Use
+
+- When MCP-based browser control is required instead of a CLI workflow.
+
 
 ## Installation
 
@@ -50,28 +59,33 @@ agent-browser get text @e1        # Get heading text
 Ref syntax: `@e1`, `e1`, or `ref=e1` all work.
 
 **Why refs?**
+
 - Deterministic: points to exact element from snapshot
 - Fast: no DOM re-query, uses cached role/name locator
 - AI-friendly: snapshot + ref workflow optimal for LLMs
 
 ### Interactive Roles (get refs automatically)
+
 `button`, `link`, `textbox`, `checkbox`, `radio`, `combobox`, `listbox`, `menuitem`, `option`, `searchbox`, `slider`, `spinbutton`, `switch`, `tab`, `treeitem`
 
 ### Content Roles (get refs when named)
+
 `heading`, `cell`, `gridcell`, `columnheader`, `rowheader`, `listitem`, `article`, `region`, `main`, `navigation`
 
 ## Commands
 
 ### Navigation
+
 ```bash
 agent-browser open <url>          # Auto-prepends https:// if needed
 agent-browser back
-agent-browser forward  
+agent-browser forward
 agent-browser reload
 agent-browser close               # Closes browser + daemon
 ```
 
 ### Interaction
+
 ```bash
 agent-browser click <sel>         # Left click
 agent-browser dblclick <sel>      # Double click
@@ -92,6 +106,7 @@ agent-browser upload <sel> <file1> [file2...]
 ```
 
 ### Get Info
+
 ```bash
 agent-browser get text <sel>      # textContent
 agent-browser get html <sel>      # innerHTML
@@ -104,6 +119,7 @@ agent-browser get box <sel>       # Bounding box {x,y,width,height}
 ```
 
 ### Check State
+
 ```bash
 agent-browser is visible <sel>
 agent-browser is enabled <sel>
@@ -111,6 +127,7 @@ agent-browser is checked <sel>
 ```
 
 ### Snapshot Options
+
 ```bash
 agent-browser snapshot            # Full accessibility tree
 agent-browser snapshot -i         # Interactive elements only (buttons, inputs, links)
@@ -121,6 +138,7 @@ agent-browser snapshot -i -c -d 5 # Combine options
 ```
 
 ### Screenshots & PDF
+
 ```bash
 agent-browser screenshot [path]   # PNG to stdout (base64) or file
 agent-browser screenshot --full   # Full page screenshot
@@ -129,6 +147,7 @@ agent-browser pdf <path>          # Save as PDF (format: Letter, A4, etc)
 ```
 
 ### Wait
+
 ```bash
 agent-browser wait <selector>     # Wait for element visible
 agent-browser wait <ms>           # Wait for time (numeric = milliseconds)
@@ -139,6 +158,7 @@ agent-browser wait --fn "window.ready === true"  # Wait for JS condition
 ```
 
 ### Find (Semantic Locators)
+
 ```bash
 agent-browser find role button click --name "Submit"
 agent-browser find text "Sign In" click
@@ -155,6 +175,7 @@ agent-browser find nth 2 "a" text          # Nth match (0-indexed)
 Actions: `click`, `fill`, `check`, `hover`, `text`
 
 ### Sessions (Isolated Browser Instances)
+
 ```bash
 agent-browser --session agent1 open site-a.com
 agent-browser --session agent2 open site-b.com
@@ -167,6 +188,7 @@ agent-browser session             # Show current session
 Each session has: own browser, cookies/storage, navigation history, auth state.
 
 ### Storage & Cookies
+
 ```bash
 agent-browser cookies             # Get all cookies
 agent-browser cookies set <name> <val>
@@ -180,6 +202,7 @@ agent-browser storage session     # Same for sessionStorage
 ```
 
 ### Tabs & Windows
+
 ```bash
 agent-browser tab                 # List tabs
 agent-browser tab new [url]       # New tab
@@ -189,18 +212,21 @@ agent-browser window new          # New window (new context)
 ```
 
 ### Frames
+
 ```bash
 agent-browser frame <sel>         # Switch to iframe by selector
 agent-browser frame main          # Back to main frame
 ```
 
 ### Dialogs (alert/confirm/prompt)
+
 ```bash
 agent-browser dialog accept [text]  # Accept (with optional prompt text)
 agent-browser dialog dismiss        # Dismiss
 ```
 
 ### Network Interception
+
 ```bash
 agent-browser network route <url>              # Intercept requests
 agent-browser network route <url> --abort      # Block requests
@@ -211,6 +237,7 @@ agent-browser network requests --filter api    # Filter by URL pattern
 ```
 
 ### Browser Settings
+
 ```bash
 agent-browser set viewport <w> <h>
 agent-browser set device "iPhone 14"   # Emulate device
@@ -222,6 +249,7 @@ agent-browser set headers '{"X-Custom":"value"}'
 ```
 
 ### Debug & Tracing
+
 ```bash
 agent-browser --headed open <url>   # Show browser window (not headless)
 agent-browser console               # View console messages
@@ -235,12 +263,14 @@ agent-browser trace stop [path]     # Stop and save trace
 ```
 
 ### Auth State Persistence
+
 ```bash
 agent-browser state save auth.json  # Save cookies, localStorage, sessionStorage
 agent-browser state load auth.json  # Load at next launch (must be at browser start)
 ```
 
 ### Mouse Control
+
 ```bash
 agent-browser mouse move <x> <y>    # Move to coordinates
 agent-browser mouse down [button]   # left|right|middle
@@ -250,15 +280,15 @@ agent-browser mouse wheel <dy> [dx] # Scroll wheel
 
 ## Options
 
-| Option | Description |
-|--------|-------------|
+| Option             | Description                                       |
+| ------------------ | ------------------------------------------------- |
 | `--session <name>` | Isolated session (or `AGENT_BROWSER_SESSION` env) |
-| `--json` | JSON output (for parsing) |
-| `--full, -f` | Full page screenshot |
-| `--headed` | Show browser window |
-| `--name, -n` | Locator name filter |
-| `--exact` | Exact text match |
-| `--debug` | Debug output |
+| `--json`           | JSON output (for parsing)                         |
+| `--full, -f`       | Full page screenshot                              |
+| `--headed`         | Show browser window                               |
+| `--name, -n`       | Locator name filter                               |
+| `--exact`          | Exact text match                                  |
+| `--debug`          | Debug output                                      |
 
 ## Selector Types
 
@@ -283,6 +313,7 @@ agent-browser is visible @e2 --json
 ## Patterns
 
 ### Login Flow
+
 ```bash
 agent-browser open https://example.com/login
 agent-browser snapshot -i
@@ -295,6 +326,7 @@ agent-browser state save auth.json  # Persist auth for later
 ```
 
 ### Form Submission
+
 ```bash
 agent-browser open https://example.com/form
 agent-browser snapshot -i --json
@@ -308,6 +340,7 @@ agent-browser wait --text "Success"
 ```
 
 ### Scraping Data
+
 ```bash
 agent-browser open https://example.com/data
 agent-browser snapshot
@@ -318,6 +351,7 @@ agent-browser eval "JSON.stringify([...document.querySelectorAll('.price')].map(
 ```
 
 ### Multi-page Navigation
+
 ```bash
 agent-browser open https://example.com
 agent-browser snapshot -i
@@ -327,6 +361,7 @@ agent-browser snapshot -i  # Fresh snapshot for new page
 ```
 
 ### Parallel Browser Sessions
+
 ```bash
 # Terminal 1
 agent-browser --session scraper1 open https://site-a.com
@@ -338,6 +373,7 @@ agent-browser --session scraper2 snapshot -i
 ```
 
 ### Mock API Response
+
 ```bash
 agent-browser network route "**/api/users" --body '[{"id":1,"name":"Mock User"}]'
 agent-browser open https://example.com  # App sees mocked data
@@ -345,6 +381,7 @@ agent-browser network unroute "**/api/users"
 ```
 
 ### Handle File Downloads
+
 ```bash
 agent-browser open https://example.com/downloads
 agent-browser snapshot -i
@@ -367,10 +404,10 @@ Default timeout: 10 seconds (Playwright default is 30s). Commands will fail afte
 
 ## Browser Support
 
-| Platform | Binary | Fallback |
-|----------|--------|----------|
-| macOS ARM64 | Native Rust | Node.js |
-| macOS x64 | Native Rust | Node.js |
-| Linux ARM64 | Native Rust | Node.js |
-| Linux x64 | Native Rust | Node.js |
-| Windows | - | Node.js |
+| Platform    | Binary      | Fallback |
+| ----------- | ----------- | -------- |
+| macOS ARM64 | Native Rust | Node.js  |
+| macOS x64   | Native Rust | Node.js  |
+| Linux ARM64 | Native Rust | Node.js  |
+| Linux x64   | Native Rust | Node.js  |
+| Windows     | -           | Node.js  |

package/dist/template/.opencode/skill/agent-teams/SKILL.md

@@ -6,11 +6,22 @@ description: >
   Covers team configuration, task distribution, coordination patterns, and best practices.
 version: "1.0.0"
 license: MIT
+tags: [agent-coordination, workflow]
+dependencies: []
 ---
 
 # Agent Teams - Multi-Agent Team Coordination
 
-Coordinate multiple agents working as a team under a lead agent. Unlike subagents (fire-and-forget), teams enable continuous coordination, shared context, and dynamic task reassignment.
+## When to Use
+
+- Parallel research, review, or competing approaches that need coordination
+- Tasks benefit from shared findings and a lead coordinating teammates
+
+## When NOT to Use
+
+- Single-agent tasks or tightly coupled edits where coordination overhead is wasteful
+- Simple parallel work that can use fire-and-forget subagents instead
+
 
 ## Overview
 

package/dist/template/.opencode/skill/augment-context-engine/SKILL.md

@@ -1,11 +1,21 @@
 ---
 name: augment-context-engine
 description: Semantic codebase search via Augment Context Engine MCP. Use when you need to understand code relationships, find relevant implementations, or explore unfamiliar codebases beyond keyword matching.
+version: 1.0.0
+tags: [research, context]
+dependencies: []
 ---
 
 # Augment Context Engine (MCP)
 
-Semantic code search powered by Augment's Context Engine. Unlike grep (text matching), this understands code meaning, relationships, and cross-file dependencies.
+## When to Use
+
+- When you need semantic search to understand code relationships or find implementations beyond keyword matching.
+
+## When NOT to Use
+
+- When exact string matching or known file paths can be handled by grep/read/LSP locally.
+
 
 ## Available Tools