@rune-kit/rune 2.29.0 → 2.30.0

package/.codex-plugin/plugin.json

@@ -1,6 +1,6 @@
 {
   "name": "rune",
-  "version": "2.29.0",
+  "version": "2.30.0",
   "description": "66-skill engineering mesh with Codex-native skills, subagents, synchronous lifecycle hooks, security gates, MCP-aware workflows, and project memory.",
   "author": {
     "name": "Rune Contributors"

package/README.md

@@ -103,11 +103,19 @@ Plus **9 domain packs** (product, sales, data-science, support, growth, media, p
 
 ---
 
-## What's New (v2.29.0 — Codex Native)
+## What's New (v2.30.0 — Tier Restored)
+
+> **v2.30.0 (2026-07-29):** Rune's model tier table had quietly stopped applying on Claude Code, and had been wrong in **23 of 66 places** for longer than that. `agents/*.md` is a hand-written parallel copy of each skill's tier; nothing generated it from `skills/` and no check compared the two, so they drifted — `cook` said opus in one file and sonnet in the other, `verification` said sonnet and haiku. `skills/*/SKILL.md` is now the single source of truth, every pair agrees, and a new **`validateAgentSync`** gate fails CI on any future drift, alongside a model-split check and a warning for a top-level `model` that cannot take effect. Where the two disagreed the tier was decided by role, with one rule overriding both old values: **a gate never runs below sonnet** — `completion-gate`, `constraint-check`, `integrity-check`, `hallucination-guard`, and `verification` exist to catch unverified claims, so running them cheaper than the agent they audit defeats the point. Separately: tiering only ever applied through spawned subagents, and on Opus 5 the harness tells the model not to spawn them unless asked — so every tier assignment had become a silent no-op. `scout` and `docs-seeker` now declare `context: fork` + `model: haiku` and route through the Skill tool instead; verified end-to-end on Opus 5 by per-model billing showing `claude-haiku-4-5`. `cook` deliberately declares no model — an orchestrator inherits the session model, because choosing Opus for a session is a deliberate act. 1,647 tests.
+
+### Previous (v2.29.1 — One-Command Update)
+
+> **v2.29.1 (2026-07-24):** v2.29.0 documented the update flow; this release automates it. New **`rune update`** command — one-shot updater for an already-configured project: `git pull --ff-only` any detected Pro/Business tier repos (env var → sibling dir, same detection as setup; a failed pull **aborts loudly**, never a silent half-update), re-runs the managed setup rewrite in place non-interactively (your installed platforms, preset, and tiers are detected from the existing hook config — no prompts), then verifies with doctor + hook drift and reminds Codex users to re-trust `/hooks` only when `.codex/hooks.json` actually changed. Flags: `--no-pull`, `--preset`, `--tier`, `--dry`. Plus: "Updating" sections in all three tier READMEs. Docs-and-CLI patch — no skill or mesh changes. 1,641 tests.
+
+#### Earlier (v2.29.0 — Codex Native)
 
 > **v2.29.0 (2026-07-23):** Codex stops being a compile target that happens to work and becomes a **first-class native runtime**. The Codex adapter now emits everything current Codex supports natively — `.agents/skills/`, project-scoped agent TOML (`.codex/agents/rune-{heavy,standard,fast}.toml`), a dedicated synchronous hook adapter targeting `.codex/hooks.json` (Codex silently skips async handlers), MCP config, and tier-aware compilation so Free/Pro/Business stacks resolve exactly as on Claude Code. The mesh validator was fixed at the root — `Calls (outbound)` is now the single authoritative edge inventory, acknowledged by the target's `Called By` — moving the canonical count from 209 to **248 connections** (same mesh, honest count; doctor now fails CI on any stale claim). Doctor also grew a cross-tier audit: Business metadata (28 pack skills, 4 orchestrators), $149 pricing, 13-platform count, and JSON-schema validation of the new `docs/config-schema.json` + hooks manifest schema. All XLabs remote-MCP references now go through `XLABS_MCP_TOKEN` — plaintext bearer values are forbidden. 1,615 tests pass.
 
-### Previous (v2.28.0 — Reasoner's Blind Spots)
+#### Earlier (v2.28.0 — Reasoner's Blind Spots)
 
 > **v2.28.0 (2026-07-22):** Completes the reasoning wave. Every addition targets one failure class: **a check that feels done because the model re-read its own work and agreed with itself.** `problem-solver` (v0.6.0) gains a **model failure-mode table** beside its human-bias table — pattern-match satisfaction, template hijack, fluent≠true, prior-as-fact, completion pressure, surface blindness — plus three tells that you are inside one right now (instant confident answer; a stated detail your draft never used; two failed attempts in the same framing). `verification` (v0.8.0) gains the **Constraint Loop** for deliverables carrying a mechanically checkable constraint on their own surface form (banned characters, exact counts, strict formats) — a class Rune had no coverage for: expand the constraint before drafting, verify with a tool, re-scan the whole artifact, ship byte-for-byte. `design` (v0.9.0) gains **render blindness** — a checklist item ticked from source is a prediction, and the imagined render is always flattering; visual items are marked 👁 and are ticked from a render or marked ASSUMED. Advisory throughout, no new skills.
 
@@ -407,6 +415,45 @@ This compiles all 66 skills into your IDE's rules format. Same knowledge, same w
 lifecycle hooks; the remaining adapters preserve the workflows and constraints in the best
 format each runtime supports.
 
+## Updating
+
+New releases are announced on [Telegram](https://t.me/xlabs_updates) and [GitHub Releases](https://github.com/rune-kit/rune/releases); see [`CHANGELOG.md`](CHANGELOG.md) for what changed.
+
+> **One-liner alternative**: `npx @rune-kit/rune@latest update` runs the whole flow below for an already-configured project — git-pulls detected Pro/Business repos, re-runs the managed setup rewrite in place (reusing your installed platforms/preset/tiers), and verifies with doctor. The platform-specific commands below remain the canonical path.
+
+### Claude Code (plugin)
+
+```bash
+claude plugin update rune
+```
+
+Restart Claude Code to load the new version, then re-run the wizard once so hooks and tier skills match the new release:
+
+```bash
+npx @rune-kit/rune@latest setup
+```
+
+### Codex / Cursor / Windsurf / other compiled platforms
+
+Re-running setup with `@latest` fetches the newest compiler and rewrites everything Rune manages **in place** — it is idempotent, and never touches your own config or the human-authored parts of `AGENTS.md`:
+
+```bash
+# same flags you installed with, e.g.:
+npx @rune-kit/rune@latest setup --here
+npx @rune-kit/rune@latest setup --here --platform codex --tier pro
+```
+
+If you use the build pipeline (`rune init` route) instead of the wizard: `npx @rune-kit/rune@latest init --force` regenerates platform files without touching your `.rune/` state.
+
+### Verify after updating
+
+```bash
+npx @rune-kit/rune@latest doctor          # compiled output + mesh integrity
+npx @rune-kit/rune@latest doctor --hooks  # hook drift vs the canonical preset
+```
+
+On Codex, review and re-trust changed hooks with `/hooks`.
+
 ## Quick Start
 
 ```bash

package/agents/audit.md

@@ -25,7 +25,7 @@ You are the **audit** skill — Rune's comprehensive project auditor.
 
 **Critical Rules:**
 - MUST complete all 8 phases (explicitly state if skipped)
-- Only findings >80% confidence (no speculation)
+- Drop a finding only on direct counter-evidence, never on uncertainty — an unconfirmable finding is reported and typed `ASSUMED` with its unchecked premise named
 - Include at least 3 positive findings
 - Delegate Phase 1 to dependency-doctor, Phase 2 to sentinel
 

package/agents/cook.md

@@ -1,7 +1,7 @@
 ---
 name: cook
 description: "Feature implementation orchestrator — handles 70% of requests. Full TDD cycle: understand → plan → test → implement → verify → commit. Use for ANY code modification (features, bugs, refactors, security)."
-model: opus
+model: sonnet
 subagent_type: general-purpose
 ---
 

package/agents/journal.md

@@ -1,7 +1,7 @@
 ---
 name: journal
 description: "Persistent state tracking across sessions — records decisions with rationale, progress, risks, ADRs to .rune/ files. Called by surgeon, deploy, rescue for cross-session continuity."
-model: sonnet
+model: haiku
 subagent_type: general-purpose
 ---
 

package/agents/reviewer.md

@@ -9,7 +9,7 @@ You are the **reviewer** subagent — a code review specialist spawned by other
 
 ## Operating Rules
 
-1. **Every finding MUST have file:line reference** — vague observations are rejected
+1. **Every finding MUST carry a verbatim evidence snippet** — vague observations are rejected, and so are line numbers produced from memory (see Evidence & Anchoring below)
 2. Check all 5 areas: correctness, security, performance, conventions, test coverage
 3. Severity levels: CRITICAL (blocks merge) → HIGH → MEDIUM → LOW
 4. Never rubber-stamp — if zero issues found, look harder (default-suspicious mindset)
@@ -17,4 +17,26 @@ You are the **reviewer** subagent — a code review specialist spawned by other
 6. Include at least 1 positive note (what's well-designed)
 7. Verdict: APPROVE / REQUEST CHANGES / NEEDS DISCUSSION
 
+## Falsification Pass (before reporting)
+
+**Falsify, not verify.** Do not filter findings by how confident you feel — you cannot calibrate that. Filter by disproof:
+
+- **DROP** only when the code you read holds **direct counter-evidence** against the finding's key claim. Discard it silently.
+- **KEEP** when the finding rests on context outside the diff that you actually read — that context is evidence.
+- **KEEP** when you can neither confirm nor disprove it. "Unsure" is not grounds to drop.
+
+Type every surviving finding `OBSERVED` (you read the defect), `DERIVED` (follows from what you read, state the mechanism), or `ASSUMED` (rests on an unverified premise — name it). **`ASSUMED` findings are never CRITICAL.**
+
+## Evidence & Anchoring
+
+Report the snippet, not the line number. A line number recalled from a long context drifts, and a correct finding pointing at the wrong line is unactionable.
+
+- **Copy** the offending lines verbatim from the file — no rewriting, no re-indenting, diff markers (`+`/`-`) stripped, max 5 lines. A finding needing more than 5 lines to show is a design comment: report it without evidence at MEDIUM or below.
+- **Anchor** each finding before reporting: `Grep` the exact snippet in its file → hit, use that line. Miss → retry once with whitespace normalised and the first and last lines dropped.
+- **`UNANCHORED`** on a second miss: downgrade one severity level, report as `path (unanchored)` with the snippet shown inline. Never drop it — a failed `Grep` is not counter-evidence.
+
+## Strict Focus
+
+Read anything you need for context. Report findings **only** about files in the stated review scope — an out-of-scope finding is dropped, or noted as a one-line follow-up in the report footer with no severity.
+
 You do NOT fix code. You identify issues with evidence. The parent skill decides next steps.

package/agents/session-bridge.md

@@ -1,7 +1,7 @@
 ---
 name: session-bridge
 description: "Cross-session context persistence — auto-saves decisions, conventions, progress, instincts to .rune/ files. Loads + integrity-checks at session start. Auto-triggered before compaction."
-model: sonnet
+model: haiku
 subagent_type: general-purpose
 ---
 

package/agents/skill-forge.md

@@ -1,7 +1,7 @@
 ---
 name: skill-forge
 description: "The skill that builds skills — TDD-driven: baseline test → write skill → verify → refactor → integrate into mesh. Use for creating or editing Rune skills."
-model: sonnet
+model: opus
 subagent_type: general-purpose
 ---
 

package/agents/skill-router.md

@@ -1,7 +1,7 @@
 ---
 name: skill-router
 description: "Meta-enforcement layer — routes EVERY agent action through the correct skill before any code is written. Always active. Prevents 'just this once' bypasses."
-model: opus
+model: sonnet
 subagent_type: general-purpose
 ---
 

package/agents/trend-scout.md

@@ -1,7 +1,7 @@
 ---
 name: trend-scout
 description: "Market intelligence — competitor analysis, technology trends, community sentiment. Scans Product Hunt, GitHub Trending, HN, Reddit. Use for positioning, not API docs."
-model: sonnet
+model: haiku
 subagent_type: general-purpose
 ---
 

package/agents/worktree.md

@@ -1,7 +1,7 @@
 ---
 name: worktree
 description: "Git worktree lifecycle — create isolated workspaces for parallel development, manage branches, cleanup. Max 3 active worktrees. Called by team for parallel streams."
-model: sonnet
+model: haiku
 subagent_type: general-purpose
 ---
 

package/compiler/__tests__/skill-attribution.test.js

@@ -0,0 +1,109 @@
+/**
+ * Attribution hygiene — skills and agents must not carry graft-source credits.
+ *
+ * Rune skills present their guidance as Rune's own. A `> From <repo> (N★): "…"`
+ * blockquote leaks where a pattern was grafted from, dates the file with a star
+ * count, and reads as borrowed authority. The prose stays; the credit line goes.
+ *
+ * Guards the patterns removed in a2df7b5 so they cannot drift back in.
+ *
+ * Legitimate credits are NOT attribution: a license obligation for redistributed
+ * data (skills/design), a descriptive "Source patterns:" lead-in, and a filename
+ * cited inside a config example all stay. Each is asserted explicitly below so a
+ * future tightening of the rules cannot silently outlaw them.
+ */
+
+import assert from 'node:assert';
+import { readdirSync, readFileSync, statSync } from 'node:fs';
+import { dirname, join, relative, resolve } from 'node:path';
+import { describe, test } from 'node:test';
+import { fileURLToPath } from 'node:url';
+
+const __dirname = dirname(fileURLToPath(import.meta.url));
+const REPO_ROOT = resolve(__dirname, '../..');
+const SCAN_DIRS = ['skills', 'agents'];
+
+/**
+ * Each rule matches a single line. Keep them shape-based: a rule that keys on a
+ * repo name rots the moment a different repo is grafted.
+ */
+const FORBIDDEN = [
+  { name: 'blockquote "> From <repo>"', re: /^\s*>\s*From\s+\S/ },
+  { name: 'blockquote "> Inspired by"', re: /^\s*>\s*Inspired by\s/ },
+  { name: 'blockquote "> Source:"', re: /^\s*>\s*Source:/ },
+  { name: 'bare "Source: … ★"', re: /^\s*Source:\s.*★/ },
+  { name: 'blockquote carrying a star count', re: /^\s*>\s.*★/ },
+];
+
+/** Lines that look adjacent to the rules above and must keep passing. */
+const MUST_PASS = [
+  'Design intelligence data from [UI/UX Pro Max](https://github.com/nextlevelbuilder/ui-ux-pro-max-skill) (MIT, 42.8k★).',
+  '> Source patterns: production-proven PostgreSQL scaling strategies.',
+  'Source: compliance-soc2.yaml',
+];
+
+function walkMarkdown(dir) {
+  const out = [];
+  for (const entry of readdirSync(dir)) {
+    const full = join(dir, entry);
+    if (statSync(full).isDirectory()) {
+      out.push(...walkMarkdown(full));
+    } else if (entry.endsWith('.md')) {
+      out.push(full);
+    }
+  }
+  return out;
+}
+
+function findViolations(text, file) {
+  const hits = [];
+  text.split(/\r?\n/).forEach((line, i) => {
+    for (const rule of FORBIDDEN) {
+      if (rule.re.test(line)) {
+        hits.push(`${file}:${i + 1} — ${rule.name} → ${line.trim()}`);
+      }
+    }
+  });
+  return hits;
+}
+
+describe('attribution hygiene', () => {
+  test('no source-attribution lines in skills/ or agents/', () => {
+    const violations = [];
+    for (const dirName of SCAN_DIRS) {
+      const dir = join(REPO_ROOT, dirName);
+      for (const file of walkMarkdown(dir)) {
+        violations.push(...findViolations(readFileSync(file, 'utf8'), relative(REPO_ROOT, file)));
+      }
+    }
+
+    assert.deepStrictEqual(
+      violations,
+      [],
+      `Attribution lines found — delete the credit, keep the prose:\n${violations.join('\n')}`,
+    );
+  });
+
+  test('rules fire on the exact lines removed in a2df7b5', () => {
+    const removed = [
+      '> From goclaw (nextlevelbuilder/goclaw, 832★): "Compact during run, not just at session boundary."',
+      '> From superpowers (obra/superpowers, 84k★): "Each fix revealing new problems elsewhere = structural issue."',
+      '> Inspired by CLI-Anything (HKUDS/CLI-Anything, 14.5k★): "Never trust exit 0."',
+      'Source: goclaw (832★) — SHA256-based loop detection distinguishes true stuck loops.',
+    ];
+
+    for (const line of removed) {
+      assert.ok(
+        FORBIDDEN.some((rule) => rule.re.test(line)),
+        `no rule caught a known attribution line: ${line}`,
+      );
+    }
+  });
+
+  test('legitimate credits and lookalikes are not flagged', () => {
+    for (const line of MUST_PASS) {
+      const caught = FORBIDDEN.filter((rule) => rule.re.test(line)).map((rule) => rule.name);
+      assert.deepStrictEqual(caught, [], `false positive on a legitimate line: ${line}`);
+    }
+  });
+});