@hanzlaa/rcode 4.3.2 → 4.4.0

package/AGENTS.md

@@ -24,7 +24,7 @@ If a user says "just keep going" or "don't stop until done", that authorization
 
 - Follow [Conventional Commits](https://www.conventionalcommits.org/) format: `type(scope): subject`
 - Types allowed: `feat`, `fix`, `docs`, `style`, `refactor`, `test`, `chore`, `perf`, `revert`
-- Scopes allowed: `agents`, `skills`, `workflows`, `templates`, `dashboard`, `docs`, `config`, `github`, `commands`, `memory`, `brand`, `cli`, `ci`, `release`, `meta`, `tasks`, `migrations`, `refs`, `state`, `hooks`, `install`, `parity`, `triggers`, `dogfood`, `namespace`, `planning`, `insights`, `help`, `roadmap`, `session`, `audits`, `execute`, `executor`, `plan`, `planner`, `readme`, `sync`, `sprint`, `agent-exp`, `extensibility`, `lens-audit`, `tiers`, `build`, `council`, `doctor`, `postinstall`, `progress`, `security`, `tools`, `uninstall`, `update`, `test`, `changelog`, `scopes`, `phases`, `references`, `kanban`, `orchestrator`, `orchpanel`, `status`, `bin`, `brain`, `dogfeed`, `new-project`, `package`, `rcode-tools`, `rihal-tools`, `team`, `usp`, `v4`, `observability`, `audit`, `agent-rules`, `cursor`, `i18n`, `phase`, `scaffold`, `campaign`, `ship`, plus numeric phase/sprint scopes (e.g. `docs(15)`, `feat(8.3)`)
+- Scopes allowed: `agents`, `skills`, `workflows`, `templates`, `dashboard`, `docs`, `config`, `github`, `commands`, `memory`, `brand`, `cli`, `ci`, `release`, `meta`, `tasks`, `migrations`, `refs`, `state`, `hooks`, `install`, `parity`, `triggers`, `dogfood`, `namespace`, `planning`, `insights`, `help`, `roadmap`, `session`, `audits`, `execute`, `executor`, `plan`, `planner`, `readme`, `sync`, `sprint`, `agent-exp`, `extensibility`, `lens-audit`, `tiers`, `build`, `council`, `doctor`, `postinstall`, `progress`, `security`, `tools`, `uninstall`, `update`, `test`, `changelog`, `scopes`, `phases`, `references`, `kanban`, `orchestrator`, `orchpanel`, `status`, `bin`, `brain`, `dogfeed`, `new-project`, `package`, `rcode-tools`, `rihal-tools`, `team`, `usp`, `v4`, `observability`, `audit`, `agent-rules`, `cursor`, `i18n`, `phase`, `scaffold`, `campaign`, `ship`, `getting-started`, plus numeric phase/sprint scopes (e.g. `docs(15)`, `feat(8.3)`)
 - Subject: lowercase first letter, imperative mood, no trailing period, under 72 chars
 - **NEVER add Claude/AI attribution to commit messages.** No "Generated with Claude Code", no "Co-Authored-By: Claude", no "🤖 Generated". The user does not want this.
 - **NEVER use `--no-verify`** to bypass hooks. If hooks fail, fix the underlying issue.

package/CONTRIBUTING.md

@@ -279,6 +279,7 @@ We use [Conventional Commits](https://www.conventionalcommits.org/) format. The
 - `templates` — memory bank or pitch templates
 - `dashboard` — Majlis/Diwan server
 - `docs` — README, METHODOLOGY, SKILLS_INDEX, USER-GUIDE, FAQ
+- `getting-started` — the `docs/getting-started.md` onboarding guide
 - `config` — team.yaml, config.yaml
 - `github` — CI/CD, issue templates, PR templates
 - `cli` — `cli/*.js` and `rcode/bin/rcode-tools.cjs`
@@ -520,7 +521,7 @@ only surface at install time.
 ### For dashboard changes
 1. `node server/dashboard.js` starts cleanly
 2. Visit `http://localhost:7717` — all sections render
-3. Stop with `kill $(lsof -t -i:7717)`
+3. Stop with `kill $(ss -ltnp 'sport = :7717' | awk 'NR>1{match($6,/pid=([0-9]+)/,m); print m[1]}')`
 
 ### For new workflows
 1. Read the workflow instructions end-to-end as if executing it

package/README.md

@@ -2,6 +2,16 @@
 
 <div dir="rtl">طريقة رحال</div>
 
+<p align="center"><img src="brand/dashboard-tour.gif" alt="rcode Diwan dashboard tour — cycling through the project cockpit (progress donut, current phase executing, milestone outlook), the specialist agent roster, the Kanban board (TODO / in-progress / done), and the phase-driven delivery view" width="860"></p>
+
+### An engineering org in a folder — and you can watch it work.
+
+*A roster of specialist agents, phase-driven delivery, and a persistent Memory Bank — running as plain files your IDE reads. Every number is reproducible with one command (`node benchmarks/facts.cjs`); marketing rounds up, this repo ships the receipts: **497 tests, 0 failing · 45 agents · 116 commands · 1 runtime dependency**.*
+
+<p align="center"><img src="brand/dashboard-team.png" alt="rcode agent roster grouped by function — Leadership (Sadiq / strategy, Waleed / CTO, Ahmed / delivery, Nasser / eng manager), Engineering (Omar, Haitham / frontend, Yousef / backend, Zayd / ML, Khalid / DevOps), Product (Hussain / PM, Mariam / marketing), each card showing its role and tools" width="860"></p>
+
+<p align="center"><sub>The <a href="docs/install.md">Diwan dashboard</a> (view-only, <code>node server/dashboard.js</code>) — your project's phases, sprints, agents, decisions, and live orchestration, read straight from the files in <code>.rcode/</code>.</sub></p>
+
 > **A curated composition of the best AI-development practices — shipped as files.** Surgical-change discipline, parallel-agent orchestration, persistent Memory Bank patterns, phase-driven planning — packaged as one workflow for Claude Code power users. No multi-agent harness. No vector DB. Your IDE keeps the methodology; the project keeps the memory.
 
 ```bash
@@ -10,10 +20,10 @@ pnpm dlx @hanzlaa/rcode install
 
 [![npm version](https://img.shields.io/npm/v/@hanzlaa/rcode)](https://www.npmjs.com/package/@hanzlaa/rcode)
 [![downloads](https://img.shields.io/npm/dw/@hanzlaa/rcode)](https://www.npmjs.com/package/@hanzlaa/rcode)
-[![CI](https://github.com/hanzlahabib/rihal-code/actions/workflows/test.yml/badge.svg)](https://github.com/hanzlahabib/rihal-code/actions/workflows/test.yml)
+[![CI](https://github.com/hanzlahabib/rcode/actions/workflows/test.yml/badge.svg)](https://github.com/hanzlahabib/rcode/actions/workflows/test.yml)
 [![License: MIT](https://img.shields.io/badge/license-MIT-blue.svg)](LICENSE)
 
-Status: `@hanzlaa/rcode` v4.0.0 on npm. 457 automated tests across 63 files, 45 agents, 116 commands, 87 skills. Actively dogfooded on real projects every week.
+Status: `@hanzlaa/rcode` v4.3.2 on npm. 497 automated tests (100% green) across 61 files · 45 agents · 116 commands · 129 workflows · 88 skills · **1 runtime dependency**. Every number here is reproducible — run `node benchmarks/facts.cjs`. Actively dogfooded on real projects every week.
 
 ---
 
@@ -31,18 +41,20 @@ That means two things:
 - **The dogfood loop is the test suite.** Every release is run against fresh projects (calories-counter RN, reelspeed services) before publish. Bugs surface as GitHub issues, get fixed, ship.
 - **The tool grows from real friction, not theory.** Half the v3.6.20 fixes came from a single dogfeed session where 3 parallel agents found 50+ real bugs in 4 hours.
 
-If you're a solo dev or small team using Claude Code (or Cursor, Gemini, VS Code), rcode gives you the **scaffolding a 10-person engineering org would have**: code review standards, sprint cadence, decision archives, onboarding context — without hiring the org.
+If you're a solo dev or small team using Claude Code (or Cursor, Codex, VS Code), rcode gives you the **scaffolding a 10-person engineering org would have**: code review standards, sprint cadence, decision archives, onboarding context — without hiring the org.
 
 ---
 
 ## What it actually is
 
+<p align="center"><img src="brand/hero-org-in-folder.png" alt="rcode is an engineering org in a folder: agents, commands, workflows, and skills as plain markdown" width="760"></p>
+
 Three layers, specialised for software delivery:
 
 | Layer | What lives here | Example |
 |-------|-----------------|---------|
 | **Memory** | `.rcode/memory/` — git-tracked markdown, lossless distillates | "We chose Postgres over Mongo because of JSON-B + RLS — see ADR-007" |
-| **Skills** | `rcode/skills/` — 87 phrase-activated playbooks | `rcode-sprint-checker` validates file/symbol refs before execute |
+| **Skills** | `rcode/skills/` — 88 phrase-activated playbooks | `rcode-sprint-checker` validates file/symbol refs before execute |
 | **Workflows** | `rcode/workflows/` — orchestrated multi-step paths | `/rcode-plan` runs research → planner → checker → confirm |
 
 Single agent navigates the structure. No LangChain, no AutoGen, no orchestrator process. Just folders the model can read.
@@ -84,10 +96,10 @@ What you won't get:
 
 I dogfood this hard, so the honest version:
 
-- **Not a chatbot wrapper.** Zero opinions about which LLM. Works with Claude Code, Cursor, Gemini, VS Code, Antigravity, Windsurf. Bring your own keys.
+- **Not a chatbot wrapper.** Zero opinions about which LLM. Works with Claude Code, Cursor, Codex, VS Code, Antigravity, Windsurf. Bring your own keys.
 - **Not a multi-agent framework.** No agent-to-agent message bus. One agent reads markdown structure and navigates it.
 - **Not a no-code tool.** You will read markdown files. You will write commit messages. You will type slash commands.
-- **Not finished.** v4 is solid for solo and small-team work. Open issues are tracked at the [issues page](https://github.com/hanzlahabib/rihal-code/issues) — most P1 bugs get fixed within 48 hours of a dogfeed run.
+- **Not finished.** v4 is solid for solo and small-team work. Open issues are tracked at the [issues page](https://github.com/hanzlahabib/rcode/issues) — most P1 bugs get fixed within 48 hours of a dogfeed run.
 - **Not replacing senior engineers.** It gives you their scaffolding (review standards, sprint hygiene, decision archives). You still need judgment for the hard calls.
 
 ---
@@ -100,13 +112,36 @@ I dogfood this hard, so the honest version:
 | **Specialist agents** | 1 generalist | Define in Python | Define in Python | 45 shipped |
 | **Install** | IDE extension | `pip install` + config | `pip install` + code | `pnpm dlx` — one command |
 | **Infrastructure** | Cloud API | Python server | Vector store + indexer | Zero — pure files |
-| **IDE lock-in** | Cursor only | Framework-specific | Framework-specific | Claude / Cursor / Gemini / VS Code / Antigravity / Windsurf |
+| **IDE lock-in** | Cursor only | Framework-specific | Framework-specific | Claude / Cursor / Codex / VS Code / Antigravity / Windsurf |
 | **Auditability** | Chat scrollback | Tracing dashboard | Tracing dashboard | `git log` |
 
 The point isn't "I beat LangChain." The point is **you don't need LangChain for software delivery**. You need a methodology that survives session resets, and a methodology lives in files.
 
 ---
 
+## By the numbers
+
+> **Marketing teams round up. This repo ships the script.** Every figure below is computed by [`benchmarks/facts.cjs`](benchmarks/facts.cjs) from files on disk and local CLI timings — no network, no LLM calls, no hand-entered numbers. Clone the repo and run it:
+>
+> ```bash
+> node benchmarks/facts.cjs
+> ```
+
+<p align="center"><img src="brand/hero-receipt.png" alt="benchmarks/facts.cjs output — 497 tests, 0 failing · 45 agents · 116 commands · 1 runtime dependency, computed from files on disk" width="720"></p>
+
+| Metric | Value | Why it's not a vanity number |
+|---|---|---|
+| **Portable methodology corpus** | **94,481 lines** of markdown | A 10-person eng org's playbooks (agents + commands + workflows + skills + references) — as files you own and grep, not a SaaS you rent. |
+| **Automated tests** | **497**, 100% passing across 61 files | The methodology is *guarded*, not vibes. CI runs them on every push; `node --test` reproduces locally in seconds. |
+| **Tested CLI engine** | **9,005 lines** (`rcode-tools.cjs` + `lib/`) | The deterministic brain — routing, state, planning math — is real code under test, not prompt soup. |
+| **Runtime dependencies** | **1** (`ws`, for the dashboard socket) | The view-only dashboard is pure Node stdlib. Almost nothing to audit, nothing to CVE-scan, nothing to break on `npm install`. |
+| **Specialist agents / commands / workflows / skills** | **45 / 116 / 129 / 88** | An entire engineering org, phrase-activated, that travels with you across Claude Code, Cursor, Codex, VS Code, Antigravity, Windsurf. |
+| **Core-op latency** | **~33 ms**, **0 LLM tokens** (best-of-7) | Intent routing, state reads, and milestone-health are deterministic *local* compute — a few ms over Node's own cold-start floor. The orchestration layer doesn't burn API tokens on bookkeeping the way pure-LLM agent frameworks do. |
+
+**The headline:** an entire software-delivery methodology — 94k lines of it — guarded by 497 tests, riding on a single runtime dependency, with an orchestration brain that costs **zero tokens** to think. Not a viral prompt. A system you can verify line by line.
+
+---
+
 ## Quickstart
 
 ```bash
@@ -117,7 +152,7 @@ pnpm dlx @hanzlaa/rcode install
 /rcode-init
 ```
 
-> **Don't have pnpm?** On Node 18+ / npm 11.x we recommend pnpm to avoid `npx` cache issues:
+> **Don't have pnpm?** We recommend pnpm to avoid peer-dependency resolution issues on npm 11.x and `npx` cache issues:
 > `npm install -g pnpm`
 
 `/rcode-init` detects your project state (fresh / existing / returning) and routes to the right first action. For a greenfield project it auto-routes to `/rcode-new-project`.
@@ -159,9 +194,9 @@ The non-goal: building yet another agent framework. There are enough. rcode stay
 
 ## Honest state of things
 
-- **v4.0.0** is the current release — Memory Bank now ships populated by default, brain-pull is live, and 50+ bug fixes landed from the most recent 3-project dogfeed run.
+- **v4.3.2** is the current release — Memory Bank now ships populated by default, brain-pull is live, and 50+ bug fixes landed from the most recent 3-project dogfeed run.
 - **Open issues**: ~50 — half are feature requests, the rest are backlog bugs ranked by severity.
-- **Test suite**: 339 automated tests across 58 files, 100% pass on every release. Coverage is structural (compliance + artifact schema + workflow behavioral), not line-coverage.
+- **Test suite**: 497 automated tests across 61 files, 100% pass on every release (run `node --test`). Coverage is structural (compliance + artifact schema + workflow behavioral), not line-coverage.
 - **Real users**: I run it on 4 projects daily. A handful of others run it on theirs. If you find a bug, file it — most P1s ship within 48 hours.
 - **Funding**: none. This is solo work. If your company wants commercial support, [email me](mailto:hanzla.dev@gmail.com).
 

package/cli/doctor.js

@@ -67,10 +67,17 @@ function checkCompliance(filePath) {
   if (!/^## Examples/m.test(content)) missing.push('Examples');
 
   // Negative-boundary clause (component 1 of the 5-component standard).
-  // Parse the frontmatter description so folded-block YAML is normalized
-  // before the regex runs — "Do\n  NOT" becomes "Do NOT" after normalization.
+  // Three valid locations, checked in order:
+  //   a) frontmatter `not-for:` array with at least one entry (canonical YAML form)
+  //   b) a boundary phrase in the frontmatter `description` field
+  //   c) a "## Do NOT use" heading or "do not use/include" in the body
+  // Bug #874: only (b) and (c) were checked, missing skills that use (a).
+  const notForField = frontmatter['not-for'];
+  const hasNotForField =
+    Array.isArray(notForField) && notForField.length > 0;
   const desc = typeof frontmatter.description === 'string' ? frontmatter.description : '';
   const hasBoundary =
+    hasNotForField ||
     NEGATIVE_BOUNDARY_RE.test(desc) ||
     /##[^\n]*\bnot\b/i.test(body) ||
     /\bdo not (use|include)\b/i.test(body);
@@ -188,7 +195,18 @@ function runPreflight(cwd, packageRoot) {
   // 6. Agent manifest drift (only if .rcode/ is initialized — indicates installed editors)
   if (fs.existsSync(rcodeDir)) {
     const editors = [];
-    if (fs.existsSync(path.join(cwd, '.claude/skills'))) editors.push('claude');
+    // Detect Claude install via flat .claude/commands/rcode-*.md (installer writes
+    // slash commands flat, not under a subdirectory) or .claude/skills as a fallback
+    // for legacy layouts. Bug #873: checking only .claude/skills missed fresh installs.
+    const claudeCommandsDir = path.join(cwd, '.claude/commands');
+    const hasClaudeFlatCommands =
+      fs.existsSync(claudeCommandsDir) &&
+      fs.readdirSync(claudeCommandsDir).some(
+        (f) => f.startsWith('rcode-') && f.endsWith('.md'),
+      );
+    if (hasClaudeFlatCommands || fs.existsSync(path.join(cwd, '.claude/skills'))) {
+      editors.push('claude');
+    }
     if (fs.existsSync(path.join(cwd, '.cursor/rules'))) editors.push('cursor');
     if (fs.existsSync(path.join(cwd, '.windsurf/rules'))) editors.push('windsurf');
     if (fs.existsSync(path.join(cwd, '.antigravity/agents'))) editors.push('antigravity');

package/cli/index.js

@@ -64,6 +64,10 @@ Usage:
   init           Alias for install
   update         Refresh skill files (backs up .rcode/ state first)
   uninstall      Remove rcode from the current project
+                 --yes      Skip confirmation prompts
+                 --purge    Remove everything: .rcode/ state, .planning/ scaffolds,
+                            .claude/agents/rules/, .cursor/ rcode files,
+                            .gitignore block, and .git/hooks/pre-commit
   remove         Alias for uninstall
   nuke           Wipe ALL rcode/rcode installs everywhere (global packages,
                  binaries, ~/.claude/* rcode artifacts, ~/.rcode/, project artifacts)
@@ -116,7 +120,7 @@ Getting started:
   rcode tiers        # see the Golden Path
   rcode set-profile  # choose model profile (quality | balanced | budget)
 
-Documentation: https://github.com/hanzlahabib/rihal-code
+Documentation: https://github.com/hanzlahabib/rcode
   `.trim());
 }
 

package/cli/install.js

@@ -158,6 +158,9 @@ const ConfigSchema = z.object({
   git: z.object({
     branching_strategy: z.string().optional(),
   }).optional(),
+  // Declared for validation only — default ('every') lives in the hook (rcode-hooks.cjs prompt-router).
+  // Install does NOT write this key; the feature stays dormant until hooks are opted into via /rcode-enable-hooks.
+  prompt_nudge: z.enum(['every', 'once-per-intent', 'when-stale', 'off']).optional(),
 }).passthrough();
 
 /**
@@ -367,7 +370,7 @@ function printInstallHeader(targetVersion) {
     pc.cyan('│') + '   ' + dim('A persistent context-brain for your editor') + '             ' + pc.cyan('│'),
     pc.cyan('│') + '                                                           ' + pc.cyan('│'),
     pc.cyan('│') + '   ' + dim('version  ') + pc.green('v' + v) + '                                          ' + pc.cyan('│'),
-    pc.cyan('│') + '   ' + dim('docs     ') + 'github.com/hanzlahabib/rihal-code               ' + pc.cyan('│'),
+    pc.cyan('│') + '   ' + dim('docs     ') + 'github.com/hanzlahabib/rcode               ' + pc.cyan('│'),
     pc.cyan('│') + '   ' + dim('by       ') + 'Hanzla Habib' + '                                    ' + pc.cyan('│'),
     pc.cyan('│') + '                                                           ' + pc.cyan('│'),
     pc.cyan('╰───────────────────────────────────────────────────────────╯'),
@@ -723,7 +726,7 @@ function seedStarterPlanning(target, projectName) {
   // planning artifacts. See issues #670 #671 #676.
   const STUB_BANNER =
     `<!-- INSTALL STUB — overwritten by /rcode-new-project. Delete this file or run\n` +
-    `     /rcode-new-project before committing. See https://github.com/hanzlahabib/rihal-code/issues/670 -->\n\n`;
+    `     /rcode-new-project before committing. See https://github.com/hanzlahabib/rcode/issues/670 -->\n\n`;
 
   fs.writeFileSync(projectPath,
     STUB_BANNER +
@@ -2111,6 +2114,13 @@ async function installInner(opts) {
   // Codex installs to .claude/ and AGENTS.md; lifecycle via rcode workflow bridge.
   if (opts.ides.includes('codex')) {
     console.log('  ' + dim('Codex → installing to .claude/ paths + AGENTS.md. Use `rcode workflow show <name>` to feed workflows to Codex.'));
+    // #908: /rcode-* slash commands in Codex are wired via the UserPromptSubmit
+    // hook in ~/.codex/hooks.json, which installNativeHomeSlashCommands() only
+    // writes on a GLOBAL install. A project-local install silently leaves Codex
+    // with no working slash commands — warn instead of implying success.
+    if (!opts.global) {
+      console.log('  ' + warn('Codex /rcode-* slash commands need a GLOBAL install — re-run with `--global` to wire the ~/.codex/hooks.json router. This project-local install does NOT enable them.'));
+    }
   }
 
   // Gemini IDE support deferred
@@ -2128,6 +2138,10 @@ async function installInner(opts) {
   if (opts.ides.includes('antigravity')) {
     console.log('  ' + warn('Antigravity install is experimental. Files land at .antigravity/rcode/{agents,commands}/.'));
     console.log('  ' + dim('If Antigravity expects a different path, adjust .rcode/config.yaml and re-run.'));
+    // #908: same as Codex — the UserPrompt hook is only wired on a global install.
+    if (!opts.global) {
+      console.log('  ' + warn('Antigravity /rcode-* slash commands need a GLOBAL install — re-run with `--global`. This project-local install does NOT wire the hook.'));
+    }
   }
 
   // Validate requested modules exist

package/cli/postinstall.js

@@ -134,7 +134,7 @@ More:
   rcode help         # all commands (grouped)
   rcode dashboard    # view-only Diwan on :7717
 
-Docs: https://github.com/hanzlahabib/rihal-code
+Docs: https://github.com/hanzlahabib/rcode
 `);
 }
 

package/cli/uninstall.js

@@ -949,17 +949,59 @@ async function runUninstall(args) {
   // Notice about what was intentionally preserved (#876 — never delete user data silently)
   const rcodeStillExists = plan.stateDir && fs.existsSync(path.join(cwd, '.rcode'));
   const planningStillExists = !opts.purge && fs.existsSync(path.join(cwd, '.planning'));
-  if (rcodeStillExists || planningStillExists) {
+  const agentsRulesStillExists = !opts.purge && fs.existsSync(path.join(cwd, '.claude/agents/rules'));
+  // Only flag .cursor/ as residue if rcode files are still inside it — the dir
+  // may exist for unrelated Cursor user config that was never touched by rcode.
+  const cursorStillExists = !opts.purge && (() => {
+    const rulesDir = path.join(cwd, '.cursor/rules');
+    if (!fs.existsSync(rulesDir)) return false;
+    try {
+      return fs.readdirSync(rulesDir).some(
+        (f) => f.startsWith('rcode-') || f === 'rcode.mdc' || f === 'rcode-method.mdc' || f === 'rcode',
+      );
+    } catch { return false; }
+  })();
+  const gitignoreBlockStillExists = !opts.purge && (() => {
+    try {
+      const gi = path.join(cwd, '.gitignore');
+      if (!fs.existsSync(gi)) return false;
+      const txt = fs.readFileSync(gi, 'utf8');
+      return txt.includes('# ===== rcode-managed gitignore block') || txt.includes('# >>> rcode >>>');
+    } catch { return false; }
+  })();
+  // Only flag pre-commit as residue if the rcode block is still present inside
+  // it — the hook file may exist for user-written scripts unrelated to rcode.
+  const preCommitStillExists = !opts.purge && (() => {
+    const hookPath = path.join(cwd, '.git/hooks/pre-commit');
+    if (!fs.existsSync(hookPath)) return false;
+    try {
+      return fs.readFileSync(hookPath, 'utf8').includes('# ===== rcode-managed pre-commit block =====');
+    } catch { return false; }
+  })();
+
+  if (rcodeStillExists || planningStillExists || agentsRulesStillExists || cursorStillExists || gitignoreBlockStillExists || preCommitStillExists) {
     console.log();
-    console.log(`ℹ  Preserved (your project data — not removed by default):`);
+    console.log(`ℹ  Preserved (use --purge to remove all of these):`);
     if (rcodeStillExists) {
-      console.log(`      .rcode/     phases, decisions, progress, config`);
-      console.log(`                  /rcode-init will detect this on reinstall`);
+      console.log(`      .rcode/                    project state, config, decisions`);
     }
     if (planningStillExists) {
-      console.log(`      .planning/  planning scaffolds (ROADMAP, STATE, PROJECT)`);
+      console.log(`      .planning/                 ROADMAP, STATE, PROJECT scaffolds`);
     }
-    console.log(`   To remove these on next uninstall: rcode uninstall --purge`);
+    if (agentsRulesStillExists) {
+      console.log(`      .claude/agents/rules/      agent rule files`);
+    }
+    if (cursorStillExists) {
+      console.log(`      .cursor/                   Cursor IDE integration`);
+    }
+    if (gitignoreBlockStillExists) {
+      console.log(`      .gitignore                 rcode-managed block still present`);
+    }
+    if (preCommitStillExists) {
+      console.log(`      .git/hooks/                pre-commit hook still present`);
+    }
+    console.log();
+    console.log(`   Full cleanup: rcode uninstall --purge`);
   }
 
   // IDE cache reload hint — Claude Code caches the slash-command list in memory.