@cairnkeep/cli 2.17.3 → 2.19.0

package/CHANGELOG.md

@@ -5,6 +5,44 @@ All notable user-facing changes are documented here. This project follows
 
 ## [Unreleased]
 
+## [2.19.0] - 2026-09-28
+
+### Added
+
+- Include oh-my-pi (OMP) in guided setup with a project launcher and explicit,
+  opt-in machine sync instructions.
+- Report every required machine sync command when multiple harnesses are
+  selected, while retaining the singular JSON field for compatibility.
+- Support OMP in opt-in Git-linked work evidence and preserve its diff-artifact
+  provenance.
+
+## [2.18.0] - 2026-09-28
+
+### Added
+
+- oh-my-pi (omp) integration: `omp/extensions/cairnkeep-memory.ts` registers
+  the cairn-memory MCP tools as native omp tools (collision-guarded), and
+  `omp/extensions/cairnkeep-capture.ts` stages durable-memory candidates at
+  session stop and surfaces them on the next session start, with a
+  `/cairn-staged` review command. The bridge runs as a node child process so
+  it also works under compiled Bun standalone binaries.
+- `cairn sync-omp` command, bash/zsh/fish completion entries, and
+  `verify-omp-mcp-bridge.mjs` / `verify-omp-capture.mjs` verification gates.
+- Smoke gates for the pi-mcp-bridge child environment: Bun node-shim PATH
+  stripping, `MCP_HTTP_PORT` isolation, and the node child bridge protocol.
+
+### Fixed
+
+- `pi-mcp-bridge` strips Bun `/tmp/bun-node-*` shim dirs from the stdio child
+  environment; grandchildren like `cairn memory-server` spawn a real node.
+- Work-evidence smoke used fixed calendar dates that aged past the retention
+  window; timestamps are relative to the run.
+
+### Documentation
+
+- `docs/harness-compatibility.md`: oh-my-pi (omp) section.
+- `docs/operating.md`: `CAIRN_PI_BRIDGE_COMMAND` / `CAIRN_PI_BRIDGE_ARGS`.
+
 ## [2.17.3] - 2026-09-03
 
 ### Security
@@ -534,7 +572,9 @@ stored memories.
 - Preserve executable permissions and Bash 3.2 portability on macOS.
 - Add backup-first uninstall and SQLite-safe memory export/import guidance.
 
-[Unreleased]: https://github.com/cairnkeep/cairnkeep/compare/v2.15.1...HEAD
+[Unreleased]: https://github.com/cairnkeep/cairnkeep/compare/v2.19.0...HEAD
+[2.19.0]: https://github.com/cairnkeep/cairnkeep/compare/v2.18.0...v2.19.0
+[2.18.0]: https://github.com/cairnkeep/cairnkeep/compare/v2.17.3...v2.18.0
 [2.15.1]: https://github.com/cairnkeep/cairnkeep/compare/v2.15.0...v2.15.1
 [2.15.0]: https://github.com/cairnkeep/cairnkeep/compare/v2.14.0...v2.15.0
 [2.14.0]: https://github.com/cairnkeep/cairnkeep/compare/v2.13.1...v2.14.0

package/CONTRIBUTING.md

@@ -43,7 +43,7 @@ endpoints. They are not part of the default offline test suite.
   Node-based CLI implementations.
 - `bin/cairn` is the installed command entry point.
 - `templates/` contains project scaffolding and managed instruction assets.
-- `claude/`, `opencode/`, `pi/`, and `kimi/` contain harness-specific operating
+- `claude/`, `opencode/`, `pi/`, `omp/`, and `kimi/` contain harness-specific operating
   assets.
 - `scripts/` contains sync, lifecycle, packaging, and contract checks.
 - `docs/` contains the versioned user and operator references indexed from

package/docs/harness-compatibility.md

@@ -19,6 +19,7 @@ Cairnkeep commands, hooks or native trajectory capture.
 | Kimi Code | Memory MCP, `AGENTS.md`, launcher, opt-in graph and cairn-work Skills | Launcher tested; remote MCP tested with Kimi Code 0.30.0; Skills contract-tested |
 | Qwen Code | Memory MCP, launcher, project `AGENTS.md` durable-context and playbook guidance | Launcher tested; stdio and remote MCP tested with Qwen Code 0.21.1 |
 | Pi | Memory MCP through maintained local stdio extension, native opt-in trajectory extension, launcher, graph and cairn-work prompts | Pi 0.84.1 validated minimum; deterministic and real bridge/lifecycle tests |
+| oh-my-pi (omp) | Memory MCP through maintained local stdio extension (Pi-successor API) | Smoke-tested against the local omp checkout with the deterministic bridge verification |
 | Codex CLI | Project-scoped memory MCP, launcher, and project `AGENTS.md` durable-context and playbook guidance | Setup and launcher contract-tested on POSIX and simulated native Windows; project trust remains operator-controlled |
 | Other MCP clients | Memory plus optional domain-knowledge and context-pack tools | Protocol-compatible; not automatically runtime-tested |
 
@@ -77,6 +78,53 @@ release, together with Node.js 22/24/26, Bash 3.2, and native Windows. The
 executable paths were distinct; both reported 0.84.1 because that was also the
 registry-current release.
 
+## oh-my-pi (omp)
+
+Select OMP in `cairn setup` to create the project launcher, then run
+`cairn sync-omp --apply` explicitly to install
+`extensions/cairnkeep-memory.ts` and `extensions/cairnkeep-capture.ts` under
+the omp agent root (default
+`~/.omp/agent`; `PI_CODING_AGENT_DIR` overrides it, matching omp's agent
+directory resolution). Use `--check` for drift. Uninstall removes only that
+owned path, backup-first.
+
+The extension is the Pi memory extension ported to omp's successor extension
+API: it supervises `cairn memory-server` as a local stdio child through the
+same bridge, discovers the effective MCP catalog dynamically, and registers
+each tool natively with omp under its bare MCP name. omp accepts the MCP input
+schema as a plain JSON Schema document and normalizes results to its
+`AgentToolResult` shape, preserving the trusted call metadata in `details`. A
+`cairn-memory` entry in the omp MCP config coexists without collision:
+config-served MCP tools are minted with an `mcp__` name prefix, so the two
+surfaces never claim the same tool name. Server feature gates, capability
+state, and least-authority profiles continue to decide which tools exist.
+
+Startup, calls, results, stderr retention, cancellation, and shutdown are
+bounded. The extension adds tools only: it does not run prompts, activate
+context-pack skills, create an autonomous loop, or add a remote transport.
+`bun scripts/verify-omp-mcp-bridge.mjs` renders, loads, and exercises the
+installed extension against a supervised memory server, including catalog
+parity, the result shape, cancellation, shutdown, orphan-freedom, and
+tool-name collision refusal.
+
+Sync also installs `extensions/cairnkeep-capture.ts`, the omp port of the
+Claude SessionEnd/SessionStart `memory-capture` + `memory-wakeup` hook pair.
+On the first `session_stop` of a main session (the event never fires for
+task/subagent sessions), it converts the persisted branch to transcript text
+(user/assistant text only, capped at 12000 characters keeping the most recent
+turns) and pipes it to a detached staging child that runs the shared
+extraction and writes candidates to `.planning/memory-staging/` — one file
+per session, retention capped at 5 — so staging completes even if the agent
+process exits first. Candidates are staged for review only: writes to AgentFS
+stay agent-gated via the `memory_write` tool. On `session_start` of the main
+session, staged candidates are surfaced through a UI notification and a
+`deliverAs: "nextTurn"` context message injected at the next user turn; the
+`/cairn-staged` command lists the staged files on demand. Both halves are
+fail-open and skip ephemeral side-turn runtimes whose extension actions are
+not initialized. `bun scripts/verify-omp-capture.mjs` exercises the guards,
+transcript shaping, staging contract, retention cap, session de-dupe, wakeup
+injection, and the skip paths.
+
 ## Kimi Code
 
 Kimi Code reads three MCP files, in increasing precedence:

package/docs/learning/CURRICULUM-MAP.md

@@ -1,6 +1,6 @@
 # Curriculum Coverage Map
 
-**Baseline:** Cairnkeep 2.17.3
+**Baseline:** Cairnkeep 2.19.0
 **Last reviewed:** 2026-08-19
 
 This map prevents the course from drifting behind the product. It assigns each
@@ -91,6 +91,7 @@ against `cairn help` so a new command cannot bypass a curriculum decision.
 | `cairn sync` | L02 |
 | `cairn sync-pi` | L13 and L23 |
 | `cairn sync-kimi` | L02 |
+| `cairn sync-omp` | L02 |
 | `cairn doctor` | L03; advanced repair in L13-L14 |
 | `cairn playbook` | L25 |
 | `cairn trajectory` | L13 |

package/docs/learning/FEATURE-GUIDE.md

@@ -1,6 +1,6 @@
 # Cairnkeep feature guide
 
-**Baseline:** Cairnkeep 2.17.3
+**Baseline:** Cairnkeep 2.19.0
 
 This is the step-back map for choosing and operating Cairnkeep features. The
 [course labs](https://github.com/cairnkeep/cairnkeep-course-labs) provide one
@@ -86,6 +86,7 @@ prerequisites for ordinary memory.
 | Context packs | Manual local or commit-pinned Git install plus project enablement | Immutable objects and atomic pointers under `CAIRN_PACK_BASE_DIR`; no background sync | Validate, inspect digest, check/apply update, disable/remove | `course-10-trust-context` |
 | OKF exchange | `cairn pack validate-okf`, `import-okf`, and preview-confirm `export-okf` | Local import/export by default; explicit file/shared-note allowlist; redacted output | Inspect diagnostics/provenance/output list; exact-digest apply; disable/remove imported pack | `course-10-trust-context` |
 | Pi local stdio memory | Select Pi during setup, then explicitly run `cairn sync-pi --apply` | Local memory-server child; dynamic effective tool catalog; annotations retained in trusted details, not a native Pi field | `cairn sync-pi --check`, `cairn doctor`, cancel one call, then observe awaited shutdown with no orphan | `course-12-guided-setup` |
+| OMP local stdio memory | Select OMP during setup, then explicitly run `cairn sync-omp --apply` | Local memory-server child; dynamic effective tool catalog; capture stages candidates for review | `cairn sync-omp --check`, `cairn doctor`, cancel one call, then verify session-stop staging and no orphan | `course-12-guided-setup` |
 
 ## Evidence, governance, and measurement
 

package/docs/learning/lessons/L00-why-cairnkeep.md

@@ -3,7 +3,7 @@
 **Status:** Ready
 **Track:** Quickstart
 **Time:** 10 minutes
-**Tested with:** Cairnkeep 2.17.3
+**Tested with:** Cairnkeep 2.19.0
 
 ## Outcome
 

package/docs/learning/lessons/L01-safe-trial.md

@@ -3,7 +3,7 @@
 **Status:** Ready
 **Track:** Quickstart
 **Time:** 15 minutes
-**Tested with:** Cairnkeep 2.17.3
+**Tested with:** Cairnkeep 2.19.0
 
 ## Outcome
 

package/docs/learning/lessons/L02-installation.md

@@ -3,7 +3,7 @@
 **Status:** Ready
 **Track:** Quickstart
 **Time:** 25 minutes
-**Tested with:** Cairnkeep 2.17.3 and Node.js 22 or newer
+**Tested with:** Cairnkeep 2.19.0 and Node.js 22 or newer
 
 ## Outcome
 

package/docs/learning/lessons/L03-first-project.md

@@ -3,7 +3,7 @@
 **Status:** Ready
 **Track:** Quickstart
 **Time:** 25 minutes
-**Tested with:** Cairnkeep 2.17.3
+**Tested with:** Cairnkeep 2.19.0
 
 ## Outcome
 

package/docs/learning/lessons/L19-validated-skill-improvement.md

@@ -3,7 +3,7 @@
 **Status:** Ready
 **Track:** Evidence and Evaluation
 **Time:** 50 minutes
-**Tested with:** Cairnkeep 2.17.3
+**Tested with:** Cairnkeep 2.19.0
 
 ## Outcome
 

package/docs/learning/lessons/L20-mcp-tool-profiles.md

@@ -1,7 +1,7 @@
 # L20 - Least-authority MCP tool profiles
 
 **Status:** Ready
-**Tested with:** Cairnkeep 2.17.3 and Node.js 22 or newer
+**Tested with:** Cairnkeep 2.19.0 and Node.js 22 or newer
 **Time:** 25 minutes
 
 ## Outcome

package/docs/learning/lessons/L21-context-packs.md

@@ -1,7 +1,7 @@
 # L21 - Immutable context packs and skill approval
 
 **Status:** Ready
-**Tested with:** Cairnkeep 2.17.3 and Node.js 22 or newer
+**Tested with:** Cairnkeep 2.19.0 and Node.js 22 or newer
 **Time:** 40 minutes
 
 ## Outcome

package/docs/learning/lessons/L22-native-windows.md

@@ -3,7 +3,7 @@
 **Status:** Ready
 **Track:** Operator
 **Time:** 30 minutes
-**Tested with:** Cairnkeep 2.17.3, Windows x64, and Node.js 22 or newer
+**Tested with:** Cairnkeep 2.19.0, Windows x64, and Node.js 22 or newer
 
 ## Outcome
 

package/docs/learning/lessons/L23-guided-setup.md

@@ -3,7 +3,7 @@
 **Status:** Ready
 **Track:** Quickstart and Operator
 **Time:** 35 minutes
-**Tested with:** Cairnkeep 2.17.3; Pi 0.84.1; Node.js 22, 24, and 26
+**Tested with:** Cairnkeep 2.19.0; Pi 0.84.1; Node.js 22, 24, and 26
 
 ## Outcome
 
@@ -13,13 +13,14 @@ Pi's local stdio memory bridge, and recover or uninstall every managed asset
 without hidden machine changes.
 
 The selected harnesses come from Cairnkeep's declarative registry. Codex is a
-supported selection: it receives `.codex/config.toml` and a launcher, while Pi
-continues to require its separate explicit machine sync.
+supported selection: it receives `.codex/config.toml` and a launcher. Pi and
+OMP receive project launchers and each require a separate, explicit machine
+sync; selecting both reports both sync commands.
 
 ## Prerequisites
 
 - Node.js 22 or newer, npm, and Git for the main exercise.
-- Pi 0.84.1 or newer.
+- Pi 0.84.1 or newer; OMP is optional.
 - A disposable parent directory containing no credentials, secrets, private
   state, or identifying project data.
 
@@ -32,7 +33,7 @@ cairn setup /tmp/cairnkeep-guided-selector-lab
 ```
 
 Choose Git with Up/Down and Enter, tick Pi with Space in the harness checklist,
-keep local memory, then inspect the plan before confirming. Pressing Escape or
+optionally tick OMP as well, keep local memory, then inspect the plan before confirming. Pressing Escape or
 Ctrl-C cancels; redirected and limited terminals use the documented text-input
 fallback. Remove that disposable selector lab before continuing if you applied
 the plan.
@@ -63,6 +64,14 @@ cairn doctor
 ./.ai/start-pi.sh
 ```
 
+If OMP was selected, run its separate sync and use the generated launcher:
+
+```bash
+cairn sync-omp --apply
+cairn sync-omp --check
+./.ai/start-omp.sh
+```
+
 In the disposable Pi session, list the registered Cairnkeep tools and make one
 harmless read-only memory call. The extension starts `cairn memory-server` as a
 local stdio child in this project; it does not select HTTP or a remote server.
@@ -107,9 +116,10 @@ cairn sync-pi --check
 cairn doctor
 ```
 
-Doctor reports project-state drift separately from missing or drifted Pi
+Doctor reports project-state drift separately from missing or drifted harness
 machine assets. Setup repairs only recorded project assets; `cairn sync-pi
---apply` repairs only the package-owned Pi extension and prompt paths.
+--apply` and `cairn sync-omp --apply` repair only their respective package-owned
+machine paths.
 
 Preview backup-first removal before confirming it:
 
@@ -128,6 +138,7 @@ and context packs remain unless their separate purge flags are supplied.
 | Setup refuses a missing target | The non-interactive command omitted an explicit choice or `--yes` | Run `cairn setup PATH` in a terminal, or supply the target, `--git`, `--harness`, `--memory`, and `--yes` together |
 | Setup refuses `--git existing` | The target is not inside a Git work tree | Choose an existing repository, or explicitly authorize `--git init`; use `--git none` only when limited mode is intended |
 | Doctor reports missing or drifted Pi assets | Project setup selected Pi but machine sync is incomplete | Run `cairn sync-pi --apply`, then `cairn sync-pi --check` and `cairn doctor` |
+| OMP is missing an extension | Project setup selected OMP but its explicit machine sync was not applied | Run `cairn sync-omp --apply`, then `cairn sync-omp --check` and `cairn doctor` |
 | A cancelled call ends the session | The Pi/bridge version does not satisfy the per-call cancellation contract | Preserve sanitized diagnostics, close the session, and do not claim release readiness |
 | Pi exits but a child process remains | Shutdown did not complete or the tested extension drifted | Preserve process-state evidence, run the sync check, and treat the release gate as failed |
 | A client expects native annotation hints | Pi's public tool API has no annotations field | Inspect trusted bridge metadata/result `details`; never claim native propagation |

package/docs/learning/lessons/L24-okf-exchange.md

@@ -1,7 +1,7 @@
 # L24 - Reviewed knowledge exchange
 
 **Status:** Ready
-**Tested with:** Cairnkeep 2.17.3 and Node.js 22 or newer
+**Tested with:** Cairnkeep 2.19.0 and Node.js 22 or newer
 **Time:** 35 minutes
 
 ## Outcome

package/docs/learning/lessons/L25-playbooks.md

@@ -1,7 +1,7 @@
 # L25 - Bounded workflow playbooks
 
 **Status:** Ready
-**Tested with:** Cairnkeep 2.17.3 and Node.js 22 or newer
+**Tested with:** Cairnkeep 2.19.0 and Node.js 22 or newer
 **Time:** 35 minutes
 
 ## Outcome

package/docs/operating.md

@@ -75,7 +75,7 @@ cairn setup /path/to/project --git init --harness claude,pi --memory local --yes
 ```
 
 The deterministic form accepts `--git init|existing|none`, a comma-separated
-subset of `claude,opencode,pi,kimi,qwen,codex`, and `--memory local|none`. `--git
+subset of `claude,opencode,pi,omp,kimi,qwen,codex`, and `--memory local|none`. `--git
 existing` requires the target to be in an existing work tree. `--git init`
 requires Git and is the only non-interactive authorization to initialize a
 repository. `--git none` is an explicit limited mode: repository-aware features
@@ -95,16 +95,21 @@ selected harnesses, and digests/modes/template identifiers for setup-owned
 assets. It contains no credentials, endpoints, or absolute paths.
 
 Setup never installs or refreshes machine-level harness assets. Its
-`machine_sync.automatic` field is always false, and the human output labels the
-reported command as not run automatically. For selections such as Codex and
-Qwen that need no machine sync, `machine_sync.command` is `null` and the human
-output says that sync is not required. Apply any reported command explicitly,
-check it, then diagnose and launch:
+`machine_sync.automatic` field is always false, and the human output labels each
+reported command as not run automatically. The JSON result's
+`machine_sync.commands` lists every required command; the older singular
+`machine_sync.command` field remains as the first command for compatibility.
+For selections such as Codex and Qwen that need no machine sync,
+`machine_sync.commands` is empty and the human output says that sync is not
+required. Apply each reported command explicitly, check the corresponding
+assets, then diagnose and launch:
 
 ```bash
 cairn sync --apply                    # Claude Code operating assets
 cairn sync-pi --apply                 # Pi extension and prompt
 cairn sync-pi --check
+cairn sync-omp --apply                # OMP extensions
+cairn sync-omp --check
 cd /path/to/project && cairn doctor
 ./.ai/start-pi.sh
 ```
@@ -461,6 +466,8 @@ vendor or host.
 | `CAIRN_LLM_API_KEY` | API key for the extraction / embeddings endpoint (unset → substring-only memory) |
 | `CAIRN_LLM_API_URL` | Base URL of the OpenAI-compatible endpoint |
 | `CAIRN_LLM_EXTRACTION_MODEL` | Chat model for `memory-capture` extraction |
+| `CAIRN_PI_BRIDGE_COMMAND` | Memory-server command for the `pi-bridge-child` stdio host (default `cairn`) |
+| `CAIRN_PI_BRIDGE_ARGS` | JSON argument array for the bridge child server command (default `["memory-server"]`) |
 | `CAIRN_MEMORY_EMBEDDING_URL` | Embeddings endpoint (falls back to `CAIRN_LLM_API_URL`) |
 | `CAIRN_MEMORY_EMBEDDING_MODEL` | Embedding model name (required for semantic search) |
 | `CAIRN_MEMORY_EMBEDDING_TIMEOUT_MS` | Embedding request timeout before substring fallback (default `15000`) |

package/examples/eval/bundled-fake.json

@@ -1 +1 @@
-{"identifier":"cairn-offline-fake-v1","package_version":"2.17.3","schema_version":1,"task_set_digest":"13e95379f35839593047ec34521a65a8fa3ab92bd3993bca9dc8bf5c4a9ff36f"}
+{"identifier":"cairn-offline-fake-v1","package_version":"2.19.0","schema_version":1,"task_set_digest":"13e95379f35839593047ec34521a65a8fa3ab92bd3993bca9dc8bf5c4a9ff36f"}

package/mcp-memory-server/dist/artifact-schema.js

@@ -32,7 +32,7 @@ export const artifactNodeRefSchema = z.object({
 export const artifactProvenanceSchema = z.object({
     producer: z.string().min(1).max(256).refine((value) => !/[\u0000-\u001f\u007f]/.test(value)),
     source_event: z.string().min(1).max(256).optional(),
-    harness: z.enum(["claude-code", "opencode", "pi"]).optional(),
+    harness: z.enum(["claude-code", "opencode", "pi", "omp"]).optional(),
     harness_version: z.string().min(1).max(128).optional(),
     native_id: z.string().min(1).max(256).optional(),
 }).strict();

package/mcp-memory-server/dist/pi-bridge-child.js

@@ -0,0 +1,109 @@
+import { childEnvironment, connectCairnPiBridge } from "./pi-mcp-bridge.js";
+const STDERR_LIMIT_BYTES = 16 * 1024;
+let stderrTail = "";
+const stderrWrite = process.stderr.write.bind(process.stderr);
+process.stderr.write = ((chunk, ...rest) => {
+    const text = typeof chunk === "string" ? chunk : String(chunk);
+    const next = stderrTail + text;
+    stderrTail = next.length <= STDERR_LIMIT_BYTES ? next : next.slice(next.length - STDERR_LIMIT_BYTES);
+    return stderrWrite(chunk, ...rest);
+});
+function send(message) {
+    process.stdout.write(`${JSON.stringify(message)}\n`);
+}
+function fatal(error) {
+    const message = error instanceof Error ? error.message : String(error);
+    process.stderr.write(`${JSON.stringify({ type: "error", error: message })}\n`);
+    process.exit(1);
+}
+process.on("uncaughtException", fatal);
+process.on("unhandledRejection", (reason) => fatal(reason));
+async function main() {
+    const cwd = process.argv[2] ?? process.cwd();
+    const command = process.env.CAIRN_PI_BRIDGE_COMMAND ?? "cairn";
+    const args = JSON.parse(process.env.CAIRN_PI_BRIDGE_ARGS ?? '["memory-server"]');
+    const bridge = await connectCairnPiBridge({ cwd, command, args, env: childEnvironment(process.env) });
+    const controllers = new Map();
+    let stdinBuffer = "";
+    let closing = false;
+    const close = async () => {
+        if (closing)
+            return;
+        closing = true;
+        await bridge.close().catch(() => undefined);
+        process.exit(0);
+    };
+    process.on("SIGTERM", () => void close());
+    process.on("SIGINT", () => void close());
+    process.stdin.on("end", () => void close());
+    async function handle(request) {
+        try {
+            switch (request.type) {
+                case "list": {
+                    send({ type: "result", id: request.id, ok: true, tools: await bridge.listAllTools() });
+                    break;
+                }
+                case "call": {
+                    const controller = new AbortController();
+                    controllers.set(request.id, controller);
+                    try {
+                        const result = await bridge.call(request.tool, request.args ?? {}, { signal: controller.signal });
+                        send({ type: "result", id: request.id, ok: true, result });
+                    }
+                    catch (error) {
+                        send({
+                            type: "result",
+                            id: request.id,
+                            ok: false,
+                            error: error instanceof Error ? error.message : String(error),
+                            stderr: stderrTail,
+                        });
+                    }
+                    finally {
+                        controllers.delete(request.id);
+                    }
+                    break;
+                }
+                case "cancel": {
+                    controllers.get(request.id)?.abort(new Error("cancelled"));
+                    break;
+                }
+                case "close": {
+                    await close();
+                    break;
+                }
+            }
+        }
+        catch (error) {
+            const id = typeof request.id === "number" ? request.id : -1;
+            send({
+                type: "result",
+                id,
+                ok: false,
+                error: error instanceof Error ? error.message : String(error),
+                stderr: stderrTail,
+            });
+        }
+    }
+    process.stdin.on("data", (chunk) => {
+        stdinBuffer += chunk.toString("utf8");
+        let newline = stdinBuffer.indexOf("\n");
+        while (newline >= 0) {
+            const line = stdinBuffer.slice(0, newline);
+            stdinBuffer = stdinBuffer.slice(newline + 1);
+            newline = stdinBuffer.indexOf("\n");
+            if (!line.trim())
+                continue;
+            let request;
+            try {
+                request = JSON.parse(line);
+            }
+            catch {
+                continue;
+            }
+            void handle(request);
+        }
+    });
+    send({ type: "ready" });
+}
+main().catch(fatal);

package/mcp-memory-server/dist/pi-mcp-bridge.js

@@ -11,11 +11,23 @@ function positiveInteger(value, fallback, name) {
         throw new Error(`${name} must be a positive integer.`);
     return selected;
 }
-function childEnvironment(source) {
+// Bun-hosted agents (pi, omp) prepend a /tmp/bun-node-* dir whose `node` is
+// Bun itself; grandchildren like `cairn memory-server` spawn real node
+// subprocesses, so those shim dirs must not shadow it.
+const BUN_NODE_SHIM = /\/tmp\/bun-node-[^/:]+$/;
+export function childEnvironment(source) {
     const env = {};
     for (const [name, value] of Object.entries(source)) {
-        if (value !== undefined && name !== "MCP_HTTP_PORT")
-            env[name] = value;
+        if (value === undefined || name === "MCP_HTTP_PORT")
+            continue;
+        if (name === "PATH") {
+            env[name] = value
+                .split(":")
+                .filter((entry) => !BUN_NODE_SHIM.test(entry))
+                .join(":");
+            continue;
+        }
+        env[name] = value;
     }
     return env;
 }

package/mcp-memory-server/dist/work-evidence-schema.js

@@ -1,6 +1,6 @@
 import { z } from "zod";
 export const WORK_EVIDENCE_SCHEMA_VERSION = 1;
-export const WORK_EVIDENCE_HARNESSES = ["claude", "opencode", "pi", "kimi", "qwen", "codex"];
+export const WORK_EVIDENCE_HARNESSES = ["claude", "opencode", "pi", "omp", "kimi", "qwen", "codex"];
 export const WORK_EVIDENCE_DEFAULT_RETENTION_DAYS = 30;
 export const WORK_EVIDENCE_DEFAULT_STORE_MAX_BYTES = 64 * 1024 * 1024;
 export const WORK_EVIDENCE_DEFAULT_MAX_TOUCHED_PATHS = 4096;

package/mcp-memory-server/dist/work-evidence-store.js

@@ -292,7 +292,7 @@ export async function finishWorkEvidence(projectRoot, evidenceId, exitStatus, op
             provenance: {
                 producer: "cairn-work-evidence",
                 source_event: "launcher-exit",
-                harness: start.harness === "claude" ? "claude-code" : ["opencode", "pi"].includes(start.harness) ? start.harness : undefined,
+                harness: start.harness === "claude" ? "claude-code" : ["opencode", "pi", "omp"].includes(start.harness) ? start.harness : undefined,
             },
             content: { text: patchCapture.text },
         }, { ...artifactLimits, artifactMaxBytes: Math.min(artifactLimits.artifactMaxBytes, limits.patchMaxBytes) });

package/mcp-memory-server/package.json

@@ -2,7 +2,7 @@
     "name": "@cairnkeep/memory-server",
     "version": "0.1.0",
     "private": true,
-    "description": "Durable, harness-agnostic memory MCP server for coding agents — the core of Cairnkeep.",
+    "description": "Durable, harness-agnostic memory MCP server for coding agents \u2014 the core of Cairnkeep.",
     "license": "Apache-2.0",
     "type": "module",
     "engines": {
@@ -61,10 +61,12 @@
         "check:playbook": "node scripts/smoke-playbook.mjs",
         "check:notes": "npm run build && npm run check:note-signatures && npm run check:note-distillation && npm run check:note-enrichment",
         "test": "npm run test:smoke",
-        "test:smoke": "npm run build && npm run check:embeddings && npm run check:extract && npm run check:scope-guard && npm run check:http-guard && npm run check:remote-context && npm run check:explore-guard && npm run check:explore-cache && npm run check:explore-crossref && npm run check:route-guard && npm run check:reviewed-memory && npm run check:memory-proposals && npm run check:trajectory-roundtrip && npm run check:trajectory-redaction && npm run check:trajectory-retention && npm run check:note-signatures && npm run check:note-distillation && npm run check:note-enrichment && npm run check:node-compat && npm run check:typed-nodes && npm run check:memory-import && npm run check:note-mcp && npm run check:node-doctor && npm run check:artifact-store && npm run check:artifact-mcp && npm run check:compaction-capture && npm run check:capability-contract && npm run check:capability-logging && npm run check:capability-harness && npm run check:capability-mcp && npm run check:eval-schema && npm run check:eval-process && npm run check:eval-statistics && npm run check:skill-lifecycle && npm run check:graph-cli && npm run check:mcp-trust && npm run check:pi-mcp-bridge && npm run check:context-pack && npm run check:context-pack-retrieval && npm run check:context-pack-v216-compat && npm run check:context-usage && npm run check:domain-retrieval && npm run check:retrieval-benchmark && npm run check:work-evidence && npm run check:okf && npm run check:playbook",
+        "test:smoke": "npm run build && npm run check:embeddings && npm run check:extract && npm run check:scope-guard && npm run check:http-guard && npm run check:remote-context && npm run check:explore-guard && npm run check:explore-cache && npm run check:explore-crossref && npm run check:route-guard && npm run check:reviewed-memory && npm run check:memory-proposals && npm run check:trajectory-roundtrip && npm run check:trajectory-redaction && npm run check:trajectory-retention && npm run check:note-signatures && npm run check:note-distillation && npm run check:note-enrichment && npm run check:node-compat && npm run check:typed-nodes && npm run check:memory-import && npm run check:note-mcp && npm run check:node-doctor && npm run check:artifact-store && npm run check:artifact-mcp && npm run check:compaction-capture && npm run check:capability-contract && npm run check:capability-logging && npm run check:capability-harness && npm run check:capability-mcp && npm run check:eval-schema && npm run check:eval-process && npm run check:eval-statistics && npm run check:skill-lifecycle && npm run check:graph-cli && npm run check:mcp-trust && npm run check:pi-mcp-bridge && npm run check:pi-bridge-env && npm run check:pi-bridge-child && npm run check:context-pack && npm run check:context-pack-retrieval && npm run check:context-pack-v216-compat && npm run check:context-usage && npm run check:domain-retrieval && npm run check:retrieval-benchmark && npm run check:work-evidence && npm run check:okf && npm run check:playbook",
         "test:search": "npm run build && npm run check:search",
         "test:memory": "npm run build && npm run check:embeddings && npm run check:extract && npm run check:search",
-        "test:all": "npm run test:memory"
+        "test:all": "npm run test:memory",
+        "check:pi-bridge-env": "node scripts/smoke-pi-bridge-env.mjs",
+        "check:pi-bridge-child": "node scripts/smoke-pi-bridge-child.mjs"
     },
     "dependencies": {
         "@modelcontextprotocol/sdk": "^1.30.0",