fix(models): retarget Composer aliases to Composer 2.5

.claude-plugin/marketplace.json

@@ -8,7 +8,7 @@
     {
       "name": "cursor",
       "source": "./plugins/cursor",
-      "description": "Delegate coding tasks from Claude Code to Cursor CLI (Composer 2 by default).",
+      "description": "Delegate coding tasks from Claude Code to Cursor CLI (Composer by default).",
       "strict": false
     }
   ]

CHANGELOG.md

@@ -4,6 +4,12 @@ All notable changes to this project will be documented in this file.
 The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.1.0/)
 and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html).
 
+## Unreleased
+
+### Fixed
+
+- **Model aliases updated for Composer 2.5.** Cursor retired the Composer 2.x ids — `cursor-agent --list-models` now lists only `composer-2.5` and `composer-2.5-fast` (verified on macOS, 2026-06-10). The `composer`, `composer-fast`, and `fast` shortcuts now resolve to `composer-2.5-fast` (was the dead `composer-2-fast`), and `composer-full` resolves to `composer-2.5` (was `composer-2`). The retired `composer-2` / `composer-2-fast` ids are kept as identity passthroughs so users on older `cursor-agent` builds aren't broken. README, the `cursor-runner` agent guidance, command/package descriptions, and tests updated to match. (#8)
+
 ## 0.3.0 — /cursor:review + codebase hardening
 
 ### Added

README.md

@@ -1,7 +1,7 @@
 # cursor-plugin-cc
 
 > **Claude plans. Cursor writes. Claude reviews.**
-> A Claude Code plugin that delegates coding _execution_ to Cursor's Composer 2 — without ever leaving the Claude Code TUI.
+> A Claude Code plugin that delegates coding _execution_ to Cursor's Composer — without ever leaving the Claude Code TUI.
 
 [![CI](https://github.com/freema/cursor-plugin-cc/actions/workflows/ci.yml/badge.svg)](https://github.com/freema/cursor-plugin-cc/actions/workflows/ci.yml)
 [![License: MIT](https://img.shields.io/badge/license-MIT-blue.svg)](./LICENSE)
@@ -22,7 +22,7 @@ claude ▸ Plan written to ~/.claude/plans/health-endpoint.md
 you ▸ /cursor:from-plan --delegate
 
 plugin ▸ wrote tasks/20260427-1830-health-endpoint.md
-       ▸ handing off to cursor-agent (composer-2-fast, --force)…
+       ▸ handing off to cursor-agent (composer-2.5-fast, --force)…
 
 cursor ▸ ✓ src/routes/health.ts          (new, 24 lines)
        ▸ ✓ src/app.ts                    (mounted route)
@@ -42,7 +42,7 @@ cursor ▸ ✓ src/routes/health.ts          (1 line changed)
 
 That's the whole loop. Claude does the **thinking** (plan, review). Cursor does the **typing** (file edits, tests). You stay in one TUI.
 
-**Why this is fast:** `composer-2-fast` is Cursor's tuned-for-CLI variant — it ships small, well-scoped changes in seconds. Claude Code spends its tokens on planning and reviewing, where thinking actually matters. Two tools, each doing what they're best at.
+**Why this is fast:** `composer-2.5-fast` is Cursor's tuned-for-CLI variant — it ships small, well-scoped changes in seconds. Claude Code spends its tokens on planning and reviewing, where thinking actually matters. Two tools, each doing what they're best at.
 
 ## Install
 
@@ -73,7 +73,7 @@ The first `/cursor:setup` run tells you if `cursor-agent` is missing or unauthen
 ### Requirements
 
 - Node.js **≥ 18.18**
-- A Cursor account — paid for Composer 2 models; free works with `--model auto` or other entitled models
+- A Cursor account — paid for Composer models; free works with `--model auto` or other entitled models
 - `cursor-agent` on your `PATH` — install via `curl https://cursor.com/install -fsS | bash`
 - `cursor-agent login` completed at least once
 
@@ -134,7 +134,7 @@ Plus a `cursor-runner` subagent you can invoke from inside Claude to delegate we
 
 ## Why this plugin
 
-Short answer: **Composer 2 is genuinely good at most day-to-day coding work** — and I don't want a pile of terminal windows to drive it. I want Claude Code to be the orchestrator for everything. The flow that keeps working for me is simple: **Claude makes the plan, Composer executes it, Claude reviews the diff.** Two tools, each doing what it is best at.
+Short answer: **Composer is genuinely good at most day-to-day coding work** — and I don't want a pile of terminal windows to drive it. I want Claude Code to be the orchestrator for everything. The flow that keeps working for me is simple: **Claude makes the plan, Composer executes it, Claude reviews the diff.** Two tools, each doing what it is best at.
 
 "Why not do the whole thing inside Cursor, then?" Claude Code has a certain magic, particularly around planning. It is not purely about the underlying model — it is the whole rig (long-context sessions, subagents, the TUI, the way tools compose) that, in my experience, only really clicks inside Claude Code.
 
@@ -156,7 +156,7 @@ Hand a coding task to `cursor-agent -p …`.
 
 | Flag                   | Default                    | Effect                                                                                                                                                                                                                                                                                                                                                         |
 | ---------------------- | -------------------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
-| `--model <id>`         | `auto` (or `$CURSOR_PLUGIN_CC_DEFAULT_MODEL`) | Aliases → real Cursor ids: `composer`/`fast` → `composer-2-fast`, `composer-2` → `composer-2`, `sonnet` → `claude-4.6-sonnet-medium`, `opus` → `claude-opus-4-7-high`, `gpt`/`codex` → `gpt-5.3-codex`, `grok` → `grok-4-20`, `gemini` → `gemini-3.1-pro`, `auto` → `auto`. Unknown ids forwarded as-is. `auto` lets Cursor pick whatever model your account is entitled to — safe if you don't have a Composer 2 seat. Run `/cursor:setup --print-models` for the live list. |
+| `--model <id>`         | `auto` (or `$CURSOR_PLUGIN_CC_DEFAULT_MODEL`) | Aliases → real Cursor ids: `composer`/`fast` → `composer-2.5-fast`, `composer-full` → `composer-2.5`, `sonnet` → `claude-4.6-sonnet-medium`, `opus` → `claude-opus-4-7-high`, `gpt`/`codex` → `gpt-5.3-codex`, `grok` → `grok-4-20`, `gemini` → `gemini-3.1-pro`, `auto` → `auto`. Unknown or retired ids (e.g. `composer-2`, `composer-2-fast`) are forwarded as-is. `auto` lets Cursor pick whatever model your account is entitled to — safe if you don't have a Composer seat. Run `/cursor:setup --print-models` for the live list. |
 | `--background`         | off                        | Detach; the command returns a job id immediately.                                                                                                                                                                                                                                                                                                              |
 | `--wait`               | on (if not `--background`) | Block until finished.                                                                                                                                                                                                                                                                                                                                          |
 | `--fresh`              | off                        | Start a brand-new Cursor session (no resume).                                                                                                                                                                                                                                                                                                                  |
@@ -365,11 +365,11 @@ The plugin codebase is English, but it does not impose a language policy on **yo
 The two-phase loop above is the concept; here is the concrete workflow that falls out of it in practice, and the one I keep reaching for:
 
 1. **Plan in Claude Code and write a task file.** Describe what you want; ask Claude to draft a _task spec_ — a markdown file with **goal**, **acceptance criteria**, **files to touch**, and **how to verify** (the same five sections the `cursor-runner` subagent enforces). Save it under `tasks/<slug>.md` in the repo.
-2. **Hand the file to Cursor.** Run `/cursor:delegate @tasks/<slug>.md implement this`. The `@path` shorthand inlines the file contents into the prompt, so Cursor gets the full spec without Claude having to re-type it. Composer 2 executes — it is genuinely fast, and a precisely defined task is usually a one-shot job.
+2. **Hand the file to Cursor.** Run `/cursor:delegate @tasks/<slug>.md implement this`. The `@path` shorthand inlines the file contents into the prompt, so Cursor gets the full spec without Claude having to re-type it. Composer executes — it is genuinely fast, and a precisely defined task is usually a one-shot job.
 3. **Back in Claude Code: review the diff.** Approve, or iterate with `/cursor:resume "fix X"` on the same thread, or escalate with `/cursor:delegate --model opus --fresh <same task file>` if Composer stalled.
 4. **Keep the task files around.** `tasks/` becomes a little log of what the repo's delegated changes looked like. Handy when you want to re-delegate a similar slice — just copy an old file, tweak the bullets.
 
-Why this works: Claude's tokens go into **planning and reviewing**, where thinking matters; Cursor's Composer 2 handles **the actual typing**, where it is cheaper and faster. The task file is the contract between the two phases — if it is sloppy, no model will save you. Writing a good one is itself a skill, and it is the only skill this workflow asks of you.
+Why this works: Claude's tokens go into **planning and reviewing**, where thinking matters; Cursor's Composer handles **the actual typing**, where it is cheaper and faster. The task file is the contract between the two phases — if it is sloppy, no model will save you. Writing a good one is itself a skill, and it is the only skill this workflow asks of you.
 
 If you want Claude to do the whole thing automatically — draft the task file AND hand it off — that is what the `cursor-runner` subagent is for. Ask it to either "draft a task file for X" (it stops there) or "implement X via Cursor" (it drafts, hands off, and reports the diff).
 

plugins/cursor/agents/cursor-runner.md

@@ -1,6 +1,6 @@
 ---
 name: cursor-runner
-description: Hand off a well-specified coding task to the Cursor CLI (`cursor-agent`) via `/cursor:delegate`. Use for small-to-medium, well-scoped changes where speed matters (default model `composer-2-fast`). Do NOT use this agent for code review, design decisions, or large refactors — those stay with the main Claude conversation.
+description: Hand off a well-specified coding task to the Cursor CLI (`cursor-agent`) via `/cursor:delegate`. Use for small-to-medium, well-scoped changes where speed matters (default model `composer-2.5-fast`). Do NOT use this agent for code review, design decisions, or large refactors — those stay with the main Claude conversation.
 tools: [Bash, Read]
 ---
 
@@ -56,9 +56,9 @@ Small slices give Cursor a tight scope, make the diff reviewable, and make failu
 
 ### 4. Pick a model
 
-Default is `composer-2-fast` — Cursor's own current default and the fastest Composer variant. Escalate only when the task warrants it:
+Default is `composer-2.5-fast` — Cursor's own current default and the fastest Composer variant. Escalate only when the task warrants it:
 
-- `composer-2` (non-fast) — quality matters slightly more than latency, but the task is still well-scoped.
+- `composer-2.5` (non-fast) — quality matters slightly more than latency, but the task is still well-scoped.
 - `sonnet` (`claude-4.6-sonnet-medium`) — more than ~5 files touched, or moderate architecture changes.
 - `opus` (`claude-opus-4-7-high`) — cross-cutting refactor, subtle correctness, or a prior `composer` run failed.
 - `gpt` / `codex` (`gpt-5.3-codex`) — only when the user explicitly asks for it.
@@ -92,13 +92,13 @@ Do not paraphrase the summary, do not rewrite the file list, do not hide the cha
 - **Do not edit files yourself.** Use `Read` only to ground the prompt you send to Cursor — never to patch code directly.
 - **Do not review Cursor's diff.** Review is the main Claude conversation's job. Your job ends when you hand back Cursor's report.
 - **Do not run `/cursor:status`, `/cursor:result`, or `/cursor:cancel` on your own.** If the main conversation wants them, it will run them itself.
-- **Do not escalate models without a reason.** `composer-2-fast` is the default for a reason (speed + cost). Escalate only when the task description itself warrants it.
+- **Do not escalate models without a reason.** `composer-2.5-fast` is the default for a reason (speed + cost). Escalate only when the task description itself warrants it.
 - **Do not impose a language policy on the target repo.** Follow whatever conventions the target repo's `AGENTS.md` / `.cursor/rules` / existing code already establishes.
 
 ## Output format
 
 Return exactly what `delegate.ts` prints. One line of your own framing is fine:
 
-> Delegated to Cursor (`composer-2-fast`). Result below.
+> Delegated to Cursor (`composer-2.5-fast`). Result below.
 
 Then Cursor's block, unedited.

plugins/cursor/commands/delegate.md

@@ -1,5 +1,5 @@
 ---
-description: Delegate a coding task to the Cursor CLI agent (Composer 2 by default).
+description: Delegate a coding task to the Cursor CLI agent (Composer by default).
 argument-hint: '[--background] [--wait] [--fresh] [--resume[=chat-id]] [--model <id>] [--cloud] [--no-force] [--timeout <sec>] <task...>'
 allowed-tools: Bash(node:*)
 ---

plugins/cursor/package.json

@@ -1,7 +1,7 @@
 {
   "name": "cursor-plugin-cc",
   "version": "0.3.0",
-  "description": "Use Cursor CLI from Claude Code to delegate coding tasks to Composer 2 and other Cursor models.",
+  "description": "Use Cursor CLI from Claude Code to delegate coding tasks to Composer and other Cursor models.",
   "type": "module",
   "license": "MIT",
   "author": {
@@ -20,7 +20,7 @@
     "cursor",
     "cursor-cli",
     "cursor-agent",
-    "composer-2",
+    "composer-2.5",
     "ai-coding",
     "agent-delegation"
   ],

plugins/cursor/plugin.json

@@ -1,7 +1,7 @@
 {
   "name": "cursor",
   "version": "0.3.0",
-  "description": "Hand off tasks from Claude Code to cursor-agent. Composer 2 optimised.",
+  "description": "Hand off tasks from Claude Code to cursor-agent. Composer-optimised.",
   "author": {
     "name": "Tomas Grasl",
     "url": "https://www.tomasgrasl.cz/"

plugins/cursor/scripts/lib/cursor.mjs

@@ -8,12 +8,17 @@ import { run } from './run.mjs';
 // rotates these over time — `/cursor:setup --print-models` shows the live
 // list for the current account. Unknown ids are passed through verbatim.
 export const MODEL_ALIASES = {
-  composer: 'composer-2-fast',
-  'composer-fast': 'composer-2-fast',
-  fast: 'composer-2-fast',
+  // Short shortcuts point at Cursor's current Composer line (2.5).
+  composer: 'composer-2.5-fast',
+  'composer-fast': 'composer-2.5-fast',
+  fast: 'composer-2.5-fast',
+  'composer-full': 'composer-2.5',
+  // Current Composer ids (identity — also documents the live names).
+  'composer-2.5-fast': 'composer-2.5-fast',
+  'composer-2.5': 'composer-2.5',
+  // Retired Composer ids kept as passthrough for older cursor-agent builds.
   'composer-2-fast': 'composer-2-fast',
   'composer-2': 'composer-2',
-  'composer-full': 'composer-2',
   'composer-1.5': 'composer-1.5',
   auto: 'auto',
   sonnet: 'claude-4.6-sonnet-medium',
@@ -42,7 +47,7 @@ export const MODEL_ALIASES = {
 };
 
 // `auto` lets Cursor pick whatever model the account is entitled to —
-// safe default for users without a paid `composer-2-fast` seat. Power users
+// safe default for users without a paid Composer seat. Power users
 // can override per-invocation via `--model <id>` or globally via the env var
 // CURSOR_PLUGIN_CC_DEFAULT_MODEL.
 export const DEFAULT_MODEL = 'auto';

plugins/cursor/tests/cursor.test.mjs

@@ -6,13 +6,13 @@ import { HAPPY_FIXTURE, STUB_BIN, makeTempHome } from './helpers.mjs';
 
 describe('buildArgs', () => {
   it('includes the expected flags by default', () => {
-    const args = buildArgs({ prompt: 'hi', model: 'composer-2' });
+    const args = buildArgs({ prompt: 'hi', model: 'composer-2.5' });
     expect(args).toContain('-p');
     expect(args).toContain('--output-format');
     expect(args).toContain('stream-json');
     expect(args).toContain('--trust');
     expect(args).toContain('--model');
-    expect(args).toContain('composer-2');
+    expect(args).toContain('composer-2.5');
     expect(args.at(-1)).toBe('hi');
   });
 
@@ -51,16 +51,24 @@ describe('resolveModel', () => {
   });
 
   it('maps aliases to real Cursor ids', () => {
-    expect(resolveModel('composer')).toBe('composer-2-fast');
-    expect(resolveModel('fast')).toBe('composer-2-fast');
-    expect(resolveModel('composer-2')).toBe('composer-2');
+    expect(resolveModel('composer')).toBe('composer-2.5-fast');
+    expect(resolveModel('composer-fast')).toBe('composer-2.5-fast');
+    expect(resolveModel('fast')).toBe('composer-2.5-fast');
+    expect(resolveModel('composer-full')).toBe('composer-2.5');
+    expect(resolveModel('composer-2.5')).toBe('composer-2.5');
+    expect(resolveModel('composer-2.5-fast')).toBe('composer-2.5-fast');
     expect(resolveModel('sonnet')).toBe('claude-4.6-sonnet-medium');
     expect(resolveModel('opus')).toBe('claude-opus-4-7-high');
     expect(resolveModel('gpt')).toBe('gpt-5.3-codex');
     expect(resolveModel('grok')).toBe('grok-4-20');
     expect(resolveModel('gemini')).toBe('gemini-3.1-pro');
   });
 
+  it('keeps retired Composer ids as passthrough for older cursor-agent builds', () => {
+    expect(resolveModel('composer-2')).toBe('composer-2');
+    expect(resolveModel('composer-2-fast')).toBe('composer-2-fast');
+  });
+
   it('defaults to auto when empty (no env override)', () => {
     delete process.env.CURSOR_PLUGIN_CC_DEFAULT_MODEL;
     expect(resolveModel(undefined)).toBe('auto');
@@ -69,7 +77,7 @@ describe('resolveModel', () => {
 
   it('honours CURSOR_PLUGIN_CC_DEFAULT_MODEL when no input is given', () => {
     process.env.CURSOR_PLUGIN_CC_DEFAULT_MODEL = 'composer';
-    expect(resolveModel(undefined)).toBe('composer-2-fast');
+    expect(resolveModel(undefined)).toBe('composer-2.5-fast');
     process.env.CURSOR_PLUGIN_CC_DEFAULT_MODEL = 'some-custom-id';
     expect(resolveModel('')).toBe('some-custom-id');
   });
@@ -107,7 +115,7 @@ describe('runHeadless against stub binary', () => {
     const logPath = `${tmp.dir}/run.ndjson`;
     const result = await runHeadless({
       prompt: 'hi',
-      model: 'composer-2',
+      model: 'composer-2.5',
       force: false,
       logPath,
       timeoutSec: 10,

plugins/cursor/tests/delegate.fg.test.mjs

@@ -48,7 +48,7 @@ describe('delegate foreground', () => {
     expect(jobs.length).toBe(1);
     const job = jobs[0];
     expect(job.status).toBe('done');
-    expect(job.model).toBe('composer-2-fast');
+    expect(job.model).toBe('composer-2.5-fast');
     expect(job.cursorChatId).toBe('chat_abc123');
     expect(job.filesTouched?.length ?? 0).toBeGreaterThan(0);
   });

plugins/cursor/tests/jobs.test.mjs

@@ -31,7 +31,7 @@ describe('jobs registry', () => {
   });
 
   it('creates, reads, and updates a job atomically', () => {
-    const job = createJob({ id: 'job1', repoPath: repo, prompt: 'do it', model: 'composer-2' });
+    const job = createJob({ id: 'job1', repoPath: repo, prompt: 'do it', model: 'composer-2.5' });
     expect(job.status).toBe('running');
     const read = readJob(repo, 'job1');
     expect(read?.prompt).toBe('do it');