feat: add OpenCode plugin with full Claude Code parity (v5.7.0)

Native TypeScript plugin for OpenCode (163K stars) delivering every
Token Optimizer capability: 7-signal dual-score quality engine,
smart compaction with mode-aware context injection, session continuity
via checkpoint restore, quality nudges, loop detection, fill warnings,
tool call fatigue detection, and a built-in dashboard.

21 source files, 2537 LOC. Ships as token-optimizer-opencode on npm.
12-agent torture room gauntlet: security, correctness, parity,
reliability, performance, architecture, TypeScript, maintainability,
testing, dashboard, and repo-forensics.

Author: alexgreensh

.claude-plugin/marketplace.json

@@ -13,7 +13,7 @@
       "name": "token-optimizer",
       "source": "./",
       "description": "Audit, fix, and monitor Claude Code context window usage. Find the ghost tokens.",
-      "version": "5.6.13",
+      "version": "5.7.0",
       "author": {
         "name": "Alex Greenshpun",
         "url": "https://linkedin.com/in/alexgreensh"

.claude-plugin/plugin.json

@@ -7,7 +7,7 @@
   },
   "homepage": "https://github.com/alexgreensh/token-optimizer",
   "repository": "https://github.com/alexgreensh/token-optimizer",
-  "version": "5.6.13",
+  "version": "5.7.0",
   "license": "PolyForm-Noncommercial-1.0.0",
   "keywords": ["token", "optimization", "context", "audit", "cost", "coach"]
 }

.codex-plugin/plugin.json

@@ -1,6 +1,6 @@
 {
   "name": "token-optimizer",
-  "version": "5.6.13",
+  "version": "5.7.0",
   "description": "Audit, monitor, and reduce Codex context waste with continuity helpers and explicit outline tools.",
   "skills": "./skills/",
   "interface": {

README.md

@@ -3,10 +3,11 @@
 </p>
 
 <p align="center">
-  <a href="https://github.com/alexgreensh/token-optimizer/releases"><img src="https://img.shields.io/badge/version-5.6.13-green" alt="Version 5.6.13"></a>
+  <a href="https://github.com/alexgreensh/token-optimizer/releases"><img src="https://img.shields.io/badge/version-5.7.0-green" alt="Version 5.6.13"></a>
   <a href="https://github.com/alexgreensh/token-optimizer/releases"><img src="https://img.shields.io/github/release-date/alexgreensh/token-optimizer?label=last%20release&color=blue" alt="Last Release"></a>
   <a href="https://github.com/alexgreensh/token-optimizer"><img src="https://img.shields.io/badge/Claude_Code-Plugin-blueviolet" alt="Claude Code Plugin"></a>
   <a href="https://github.com/alexgreensh/token-optimizer/tree/main/openclaw"><img src="https://img.shields.io/badge/OpenClaw-v2.4.1-brightgreen" alt="OpenClaw v2.4.1"></a>
+  <a href="https://github.com/alexgreensh/token-optimizer/tree/main/opencode"><img src="https://img.shields.io/badge/OpenCode-v1.0.0-58a6ff" alt="OpenCode v1.0.0"></a>
   <a href="https://github.com/alexgreensh/token-optimizer/blob/main/docs/codex-beta.md"><img src="https://img.shields.io/badge/Codex-v0.1.0--beta-orange" alt="Codex v0.1.0-beta"></a>
 </p>
 <p align="center">
@@ -42,7 +43,7 @@ They compress command output, which covers 15-25% of your context on a good day.
 Token Optimizer covers all of it, keeps your work alive across compactions, measures whether the optimization actually helped, and gives you a <strong>live dashboard</strong> that shows every token, every dollar, and every turn, auto-updated after every session. Runs fully local. Zero context tokens used. Zero runtime dependencies.
 </p>
 <p align="center">
-Works on <strong>Claude Code</strong>, <strong>OpenClaw</strong>, and <strong>Codex</strong> (beta) today. Windsurf, Cursor, and more on the way.
+Works on <strong>Claude Code</strong>, <strong>OpenCode</strong>, <strong>OpenClaw</strong>, and <strong>Codex</strong> (beta) today. Windsurf, Cursor, and more on the way.
 </p>
 
 <p align="center">
@@ -93,7 +94,7 @@ git clone https://github.com/alexgreensh/token-optimizer.git ~/.claude/token-opt
 bash ~/.claude/token-optimizer/install.sh
 ```
 
-Works on Claude Code and [OpenClaw](#openclaw-plugin). Each platform has its own native plugin (Python for Claude Code, TypeScript for OpenClaw). No bridging, no shared runtime, zero cross-platform dependencies.
+Works on Claude Code, [OpenCode](#opencode), and [OpenClaw](#openclaw). Each platform has its own native plugin (Python for Claude Code, TypeScript for OpenCode and OpenClaw). No bridging, no shared runtime, zero cross-platform dependencies.
 
 </details>
 
@@ -123,6 +124,33 @@ See [`docs/codex-beta.md`](docs/codex-beta.md) for the full feature parity table
 
 </details>
 
+<details>
+<summary><h3>OpenCode</h3></summary>
+
+Native TypeScript plugin for [OpenCode](https://github.com/anomalyco/opencode) with full Claude Code feature parity. 7-signal dual-score quality engine, smart compaction with mode-aware context injection, session continuity, quality nudges, loop detection, and a built-in dashboard.
+
+```bash
+opencode plugin add token-optimizer-opencode
+```
+
+Or add to `.opencode/opencode.jsonc`:
+
+```jsonc
+{
+  "plugin": ["token-optimizer-opencode"]
+}
+```
+
+Two custom tools are available inside OpenCode:
+- `token_status` for on-demand quality reports
+- `token_dashboard` to generate and open the visual dashboard
+
+Works with every model OpenCode supports: Anthropic, OpenAI, Google, DeepSeek, Qwen, Mistral, xAI, and local models. MRCR quality curves are calibrated per model family.
+
+See [`opencode/README.md`](opencode/README.md) for full docs, configuration options, and environment variable overrides.
+
+</details>
+
 <details>
 <summary><h3>OpenClaw</h3></summary>
 
@@ -192,7 +220,7 @@ Token Optimizer handles both. And because it also checkpoints your session befor
 
 ### Fully local, zero dependencies, zero telemetry
 
-Pure Python stdlib on Claude Code and Codex. Pure Node stdlib on OpenClaw. Nothing to `pip install`, nothing to `npm install` at runtime, no analytics endpoint, no phone-home. Every measurement is a local SQLite write to a file you own under your runtime home, such as `~/.claude/_backups/token-optimizer/trends.db` or `~/.codex/_backups/token-optimizer/trends.db`. You can inspect it, export it, or delete it.
+Pure Python stdlib on Claude Code and Codex. TypeScript with zero runtime deps on OpenCode and OpenClaw. Nothing to `pip install`, no analytics endpoint, no phone-home. Every measurement is a local SQLite write to a file you own under your runtime home. You can inspect it, export it, or delete it.
 
 ### Zero context tokens consumed
 
@@ -254,7 +282,7 @@ No. All hooks are non-blocking with fail-open design. If a Token Optimizer scrip
 <details>
 <summary>📦 <strong>Does it have any runtime dependencies?</strong></summary>
 
-No. Pure Python stdlib on Claude Code and Codex. Pure Node stdlib on OpenClaw. Nothing to `pip install`, nothing to `npm install` at runtime. What you clone is everything it needs.
+No. Pure Python stdlib on Claude Code and Codex. TypeScript with zero runtime deps on OpenCode and OpenClaw. What you clone (or `opencode plugin add`) is everything it needs.
 </details>
 
 <details>

opencode/.forensicsignore

@@ -0,0 +1,5 @@
+# Bun lockfile contains SHA-512 integrity hashes (not payloads)
+bun.lock
+
+# bun:sqlite is a Bun built-in module, not an npm phantom dependency
+# (equivalent to node:fs - part of the runtime)

opencode/README.md

@@ -0,0 +1,105 @@
+# Token Optimizer for OpenCode
+
+Context quality scoring, smart compaction, and session continuity for [OpenCode](https://github.com/anomalyco/opencode). Full parity with the Claude Code Token Optimizer plugin.
+
+## What It Does
+
+Token Optimizer monitors your OpenCode sessions and helps you get the most out of your context window:
+
+- **7-signal quality scoring** with dual ResourceHealth (monotonic) + SessionEfficiency (rolling window) architecture
+- **Smart compaction** with mode-aware PRESERVE/DROP guidance (code/debug/review/infra/general)
+- **Session continuity** that restores context from prior sessions via keyword matching
+- **Quality nudges** that warn when context health drops, fill exceeds thresholds, or retry loops are detected
+- **Dashboard** with quality trends, session history, and daily aggregates
+- **`token_status` tool** for on-demand quality reports
+- **`token_dashboard` tool** to generate and open the visual dashboard
+
+## Install
+
+```bash
+opencode plugin add token-optimizer-opencode
+```
+
+## Configure
+
+Add plugin options in `.opencode/opencode.jsonc`:
+
+```jsonc
+{
+  "plugin": [
+    ["token-optimizer-opencode", {
+      "qualityWindow": 20,
+      "features": {
+        "qualityNudges": true,
+        "loopDetection": true,
+        "smartCompaction": true,
+        "continuity": true,
+        "activityTracking": true,
+        "trends": true
+      }
+    }]
+  ]
+}
+```
+
+All options are optional. Defaults are shown above.
+
+## Environment Variables
+
+Override any threshold via environment variables:
+
+| Variable | Default | Description |
+|----------|---------|-------------|
+| `TOKEN_OPTIMIZER_QUALITY_WINDOW` | 20 | Rolling window size for ratio signals |
+| `TOKEN_OPTIMIZER_TOOL_CALL_WARN` | auto | Tool call warning threshold (scales with context window) |
+| `TOKEN_OPTIMIZER_TOOL_CALL_CRITICAL` | auto | Tool call critical threshold |
+| `TOKEN_OPTIMIZER_CHECKPOINT_RETENTION_DAYS` | 7 | Days to keep checkpoints |
+| `TOKEN_OPTIMIZER_CHECKPOINT_RETENTION_MAX` | 50 | Max checkpoints to scan for restore |
+| `TOKEN_OPTIMIZER_RELEVANCE_THRESHOLD` | 0.3 | Min relevance score for checkpoint restore |
+| `TOKEN_OPTIMIZER_NUDGES` | true | Enable quality nudges |
+| `TOKEN_OPTIMIZER_LOOP_DETECTION` | true | Enable retry loop detection |
+| `TOKEN_OPTIMIZER_SMART_COMPACTION` | true | Enable compaction context injection |
+| `TOKEN_OPTIMIZER_CONTINUITY` | true | Enable session continuity |
+| `TOKEN_OPTIMIZER_ACTIVITY` | true | Enable activity tracking |
+| `TOKEN_OPTIMIZER_TRENDS` | true | Enable trends collection |
+
+## Quality Scoring
+
+The quality score uses a dual-composite architecture:
+
+**ResourceHealth** (monotonic, can only decrease within a session):
+- Context fill degradation (50%) - MRCR-curve-based quality estimate
+- Compaction depth (30%) - information loss from repeated compaction
+- Absolute waste tokens (20%) - stale reads + bloated results
+
+**SessionEfficiency** (rolling window, can rise or fall):
+- Stale reads (30%) - re-reading files after writing them
+- Bloated results (30%) - large tool outputs never referenced
+- Decision density (20%) - ratio of substantive messages
+- Agent efficiency (20%) - agent dispatch result/prompt ratio
+
+Grades: S (90+), A (80+), B (70+), C (55+), D (40+), F (<40)
+
+## Hooks Used
+
+| Hook | Purpose |
+|------|---------|
+| `chat.message` | Track user messages, trigger quality scoring |
+| `tool.execute.before` | Record file reads |
+| `tool.execute.after` | Record tool results, file writes, agent dispatches, activity tracking |
+| `experimental.chat.system.transform` | Inject warnings, restore session continuity |
+| `experimental.session.compacting` | Inject mode-aware compaction guidance, capture checkpoint |
+| `experimental.compaction.autocontinue` | Reset signals post-compaction, refresh quality |
+| `event` | Handle session lifecycle (created/deleted) |
+
+## Model Support
+
+Context window sizes are mapped for 30+ models across all major providers:
+Anthropic (Opus/Sonnet 1M, Haiku 200K), OpenAI (GPT-5.x, GPT-4.1, o3/o4), Google (Gemini 2.x/3.x),
+DeepSeek, Qwen, Mistral, xAI Grok, and more.
+
+MRCR quality curves are calibrated per model family for accurate fill-degradation estimates.
+
+## License
+
+PolyForm Noncommercial 1.0.0