release: stage v4.3.2 'Proof' (patch) — ship the post-4.3.1 fixes

npm 4.3.1 was published from the staging snapshot (2026-06-05 22:00Z) and missed
the 10 commits made right after it — including the broken-MCP-config README fix.
4.3.1 can't be republished, so this cuts 4.3.2: version bumps (package.json,
server.json x2), README banner/install/badge (1149) + a v4.3.2 what's-new note,
CHANGELOG [4.3.2], docs/release-notes-4.3.2.md.

Bundles: the MCP setup command fix (engramx serve -> engram-serve), gen-ccs
fallback, TS interface/type extraction, git-miner self-edge, provider unref, the
CI ripgrep fix, and llms-install.md. tsc clean; build ok; full suite 1149/1149;
versions consistent; tarball 55 files.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>

Author: NickCirv

CHANGELOG.md

@@ -6,6 +6,41 @@ All notable changes to engram are documented here. Format based on
 
 ## [Unreleased]
 
+## [4.3.2] — 2026-06-06 — "Proof" (patch)
+
+**Why:** 4.3.1 shipped from a staging snapshot and missed a batch of fixes made
+right after — most importantly a **broken MCP-setup command in the README**.
+This patch ships them. All fixes, no new features, no change to the honest claim.
+1149 tests.
+
+### Fixed
+- **MCP setup (README):** the documented config told MCP clients (Cline, Cursor,
+  Claude Desktop) to run `npx -y engramx serve <path>` — but there is no `serve`
+  subcommand (cli has `server` = the HTTP API; the MCP stdio server is the
+  separate `engram-serve` bin), so the server failed to start with
+  `error: unknown command 'serve'`. Corrected to `npx -y -p engramx engram-serve
+  <path>` (verified against the MCP initialize handshake).
+- **`gen-ccs`:** on a fresh repo it wrote an empty file and reported a false
+  "✅ Generated"; now falls back to the top code entities by graph degree (and
+  warns honestly when there is genuinely nothing to export).
+- **AST extraction:** TS `interface` / `type` alias / `enum` declarations are now
+  extracted as first-class nodes (kinds `interface`/`type`) instead of being
+  dropped — they were dead reference targets before.
+- **git co-change miner:** no longer emits a self-edge when two distinct files
+  share a basename stem (e.g. `src/index.ts` + `src/miners/index.ts`).
+- **Providers:** the mempalace/context7 warmup + availability probes now `unref`
+  their child processes, so a slow/absent backend can't keep the engram process
+  alive for seconds after a session's context bundle was already delivered.
+
+### Added
+- `llms-install.md` — an agent-readable setup guide (global install, `engram
+  init`, the `mcpServers` config block) so tools like Cline can configure engram
+  autonomously.
+
+### Internal
+- CI installs ripgrep (best-effort) so the grep never-worse-gate tests run on the
+  runners; first dedicated tests for the git-miner and the refs cache.
+
 ## [4.3.1] — 2026-06-06 — "Proof" (patch)
 
 **Why:** a two-pass execution-swarm audit (agents that *ran* engram end-to-end, not just read it)

README.md

@@ -1,5 +1,5 @@
 <p align="center">
-  <img src="assets/banner-v3.png" alt="engramx — the universal context spine for AI coding tools (v4.3.1 'Proof')" width="100%">
+  <img src="assets/banner-v3.png" alt="engramx — the universal context spine for AI coding tools (v4.3.2 'Proof')" width="100%">
 </p>
 
 <p align="center">
@@ -58,7 +58,7 @@
   <a href="https://open-vsx.org/extension/nickcirv/engram-vscode"><img src="https://img.shields.io/open-vsx/v/nickcirv/engram-vscode?color=blue&label=OpenVSX" alt="OpenVSX engram-vscode"></a>
   <img src="https://img.shields.io/badge/license-Apache%202.0-blue" alt="License">
   <img src="https://img.shields.io/badge/node-%3E%3D20-brightgreen" alt="Node">
-  <img src="https://img.shields.io/badge/tests-1144%20passing-brightgreen" alt="Tests">
+  <img src="https://img.shields.io/badge/tests-1149%20passing-brightgreen" alt="Tests">
   <img src="https://img.shields.io/badge/per--file%20reduction-53--89%25-orange" alt="53-89% per-file structural reduction">
   <img src="https://img.shields.io/badge/native%20deps-zero-green" alt="Zero native deps">
   <img src="https://img.shields.io/badge/LLM%20cost-$0-green" alt="Zero LLM cost">
@@ -86,8 +86,8 @@ mkdir -p /tmp/engram-demo && cd /tmp/engram-demo && \
   echo "export const buggy = () => null;" > src.ts && \
   git add -A && git commit -q -m "feat: add buggy helper returning null causing form crashes" && \
   git revert --no-edit HEAD > /dev/null && \
-  npx --yes engramx@4.3.1 init . && \
-  npx --yes engramx@4.3.1 mistakes
+  npx --yes engramx@4.3.2 init . && \
+  npx --yes engramx@4.3.2 mistakes
 ```
 
 You should see, within 30 seconds, the **bi-temporal pre-mortem** engram auto-captured from your revert:
@@ -121,9 +121,11 @@ Works in 8 IDEs and counting — Claude Code, Cursor, Cline, Continue.dev, Aider
 
 > **v4.3 "Proof" shipped 2026-06-05 — engram's saving is now real *and provable*.** Run **`engram measure`** in your own repo to see the honest structural context-token reduction on **your** code — every disclosure computed live: it's a ceiling, here's the recall, here's the intercept rate, "structural tokens, not your bill." Run **`npm run bench:recall`** for the reproducible proof that engram surfaces the files a change actually touches (recall@10 33% on engram's own repo, decomposed honestly: candidate generation reaches 43%, the PageRank ranker adds +3.2pp over random-within-candidate; 10.4% blind chance).
 >
-> New this release: a **never-worse gate** on Grep (engram passes through whenever its packet isn't actually smaller — sized to your grep's exact `cwd`/`path`/`glob` scope), **Bash-grep interception** (the shell-only IDEs — Aider, Codex CLI, Cline — now get the call-site packet too), a **sub-agent context broker** (a tight ~100-token ranked slice into each spawned Claude Code sub-agent — the one regime prompt caching can't help), and a **compaction ledger** (a "previously read" list injected at `/compact` so the agent doesn't re-explore). Every number is a measured fact or a labelled bet — **no cost claims**; a *structural* context reduction, not a bill saving (engram's net over prompt caching ≈ 0). 1144 tests. See [CHANGELOG.md](CHANGELOG.md), `docs/COMPARISON.md` (vs the other local code-graph tools — ranking isn't unique, the *combination* is), and `docs/FRONTIER.md`.
+> New this release: a **never-worse gate** on Grep (engram passes through whenever its packet isn't actually smaller — sized to your grep's exact `cwd`/`path`/`glob` scope), **Bash-grep interception** (the shell-only IDEs — Aider, Codex CLI, Cline — now get the call-site packet too), a **sub-agent context broker** (a tight ~100-token ranked slice into each spawned Claude Code sub-agent — the one regime prompt caching can't help), and a **compaction ledger** (a "previously read" list injected at `/compact` so the agent doesn't re-explore). Every number is a measured fact or a labelled bet — **no cost claims**; a *structural* context reduction, not a bill saving (engram's net over prompt caching ≈ 0). 1149 tests. See [CHANGELOG.md](CHANGELOG.md), `docs/COMPARISON.md` (vs the other local code-graph tools — ranking isn't unique, the *combination* is), and `docs/FRONTIER.md`.
 >
 > **v4.3.1 (patch) — 2026-06-06:** a two-swarm execution audit hardened the release. Fixes: the `engram cost` digest no longer prints an unqualified dollar/"% saved" figure (it now carries the structural-not-a-bill-saving caveat); `bench`/`stats` state "Nx LARGER — passes through" instead of a backwards "0.2x smaller" on tiny repos; the incremental reindex now rebuilds cross-file `calls` edges (no graph drift between full re-indexes) **and** does so in ~1ms via an mtime-keyed refs cache instead of re-parsing every file on each edit; the MCP server reports its real version; `doctor` flags a disabled kill-switch; `gen --task <bad>` errors cleanly. No new features, no behaviour change to the honest claim.
+>
+> **v4.3.2 (patch) — 2026-06-06:** **fixes a broken MCP-setup command in this README** — the config block now uses `npx -y -p engramx engram-serve <path>` (the previous `engramx serve` had no such subcommand, so MCP clients failed with "unknown command 'serve'"). Also: `gen-ccs` now exports a structural fallback on fresh repos instead of an empty file; TS `interface`/`type`/`enum` declarations are extracted as first-class graph nodes; the git co-change miner no longer emits a self-edge for distinct files sharing a basename; provider warmup probes no longer keep the process alive a few seconds after a session starts. Adds `llms-install.md` for agent-driven setup. No behaviour change to the honest claim.
 
 <details>
 <summary><strong>Earlier release notes (v4.2 "Loop", June 3)</strong></summary>

docs/release-notes-4.3.2.md

@@ -0,0 +1,42 @@
+# engramx v4.3.2 — "Proof" (patch)
+
+v4.3.1 was published from a staging snapshot and missed the batch of fixes made
+immediately after it — most importantly **a broken MCP-setup command in the
+README**. v4.3.2 ships them. All fixes, no new features, no change to the honest
+claim. 1149 tests.
+
+## Fixed
+
+- **MCP setup (README) — the important one.** The documented config told MCP
+  clients (Cline, Cursor, Claude Desktop) to run `npx -y engramx serve <path>`,
+  but there is no `serve` subcommand — the CLI has `server` (the HTTP API) and
+  the MCP stdio server is the separate `engram-serve` bin. Following the README
+  produced `error: unknown command 'serve'` and the server never started. Now
+  `npx -y -p engramx engram-serve <path>`, verified against the MCP `initialize`
+  handshake. (The other integration docs already used the correct form.)
+- **`gen-ccs`** on a fresh repo wrote an empty file under a false "✅ Generated";
+  now falls back to the top code entities by graph degree, and warns honestly
+  when there is nothing to export.
+- **AST extraction:** TS `interface` / `type` alias / `enum` declarations are now
+  first-class graph nodes (kinds `interface`/`type`) instead of being dropped —
+  they were dead reference targets.
+- **git co-change miner** no longer emits a self-edge when two distinct files
+  share a basename stem (`src/index.ts` + `src/miners/index.ts`).
+- **Providers:** mempalace/context7 warmup + availability probes `unref` their
+  child processes, so a slow/absent backend can't hold the engram process open
+  for seconds after a session's context bundle was delivered.
+
+## Added
+
+- `llms-install.md` — an agent-readable setup guide so tools like Cline can
+  configure engram autonomously (global install → `engram init` → the
+  `mcpServers` config block).
+
+## Internal
+
+- CI installs ripgrep (best-effort) so the grep never-worse-gate tests run on the
+  runners; first dedicated tests for the git-miner and the mtime-keyed refs cache.
+
+**Upgrade:** `npm i -g engramx@4.3.2` (or `npx engramx@4.3.2`). No migration.
+If you set engram up as an MCP server from the old README, update the command to
+`npx -y -p engramx engram-serve <absolute-project-path>`.

package.json

@@ -1,6 +1,6 @@
 {
   "name": "engramx",
-  "version": "4.3.1",
+  "version": "4.3.2",
   "mcpName": "io.github.NickCirv/engram",
   "description": "The context spine for AI coding agents. 9 built-in providers + mcpConfig plugin contract (wrap any MCP server in 10 lines), generic MCP-client aggregator (stdio), pre-mortem mistake-guard, bi-temporal mistake memory, Anthropic Auto-Memory bridge, SSE streaming context packets, dual-emit AGENTS.md+CLAUDE.md. Up to ~89% per-file structural context reduction (codebase-dependent; reproducible bench included). Local SQLite, zero cloud.",
   "repository": {

server.json

@@ -6,12 +6,12 @@
     "url": "https://github.com/NickCirv/engram",
     "source": "github"
   },
-  "version": "4.3.1",
+  "version": "4.3.2",
   "packages": [
     {
       "registryType": "npm",
       "identifier": "engramx",
-      "version": "4.3.1",
+      "version": "4.3.2",
       "transport": {
         "type": "stdio"
       },