@gmickel/gno 1.45.1 → 2.0.0

package/src/cli/commands/embed.ts

@@ -22,6 +22,7 @@ import {
   type WriteLeaseContention,
   withCliWriteLease,
 } from "../../core/write-lease";
+import { embedBacklog, prepareEmbeddingBacklog } from "../../embed/backlog";
 import { getEmbeddingFingerprint } from "../../embed/fingerprint";
 import {
   addUniqueSamples,
@@ -38,6 +39,7 @@ import {
   markIndexStageRunning,
   readIndexStageState,
 } from "../../embed/stage-state";
+import { countVariantBacklog } from "../../embed/variant-plan";
 import { LlmAdapter } from "../../llm/nodeLlamaCpp/adapter";
 import { resolveDownloadPolicy } from "../../llm/policy";
 import { resolveModelUri } from "../../llm/registry";
@@ -139,18 +141,6 @@ function isDisposedBatchError(message: string): boolean {
   return message.toLowerCase().includes("object is disposed");
 }
 
-async function checkVecAvailable(
-  db: import("bun:sqlite").Database
-): Promise<boolean> {
-  try {
-    const sqliteVec = await import("sqlite-vec");
-    sqliteVec.load(db);
-    return true;
-  } catch {
-    return false;
-  }
-}
-
 interface BatchContext {
   db: import("bun:sqlite").Database;
   stats: VectorStatsPort;
@@ -506,28 +496,6 @@ export async function embed(options: EmbedOptions = {}): Promise<EmbedResult> {
         const stats: VectorStatsPort = createVectorStatsPort(db);
 
         let totalToEmbed = 0;
-        if (force) {
-          const forceCount = await getActiveChunkCount(db, options.collection);
-          if (!forceCount.ok) {
-            return { success: false, error: forceCount.error.message };
-          }
-          totalToEmbed = forceCount.value;
-
-          if (totalToEmbed === 0 || dryRun) {
-            const vecAvailable = await checkVecAvailable(db);
-            return {
-              success: true,
-              embedded: totalToEmbed,
-              errors: 0,
-              contentionErrors: 0,
-              duration: 0,
-              model: modelUri,
-              searchAvailable: vecAvailable,
-              errorSamples: [],
-            };
-          }
-        }
-
         // Create LLM adapter and embedding port with auto-download
         let offline = options.offline;
         if (offline === undefined) {
@@ -585,12 +553,16 @@ export async function embed(options: EmbedOptions = {}): Promise<EmbedResult> {
           process.stderr.write("\n");
         }
 
-        // Discover dimensions via probe embedding
-        const probeResult = await embedPort.embed("dimension probe");
-        if (!probeResult.ok) {
-          return { success: false, error: probeResult.error.message };
+        const initializedPort = await embedPort.init();
+        if (!initializedPort.ok)
+          return { success: false, error: initializedPort.error.message };
+        let dimensions = embedPort.dimensions();
+        if (!embedPort.getIdentity?.()) {
+          const probeResult = await embedPort.embed("dimension probe");
+          if (!probeResult.ok)
+            return { success: false, error: probeResult.error.message };
+          dimensions = probeResult.value.length;
         }
-        const dimensions = probeResult.value.length;
 
         // Create vector index port
         const vectorResult = await createVectorIndexPort(db, {
@@ -602,6 +574,70 @@ export async function embed(options: EmbedOptions = {}): Promise<EmbedResult> {
         }
         vectorIndex = vectorResult.value;
 
+        const prepared = await prepareEmbeddingBacklog({
+          statsPort: stats,
+          embedPort,
+          vectorIndex,
+          modelUri,
+          collection: options.collection,
+          batchSize,
+          force,
+        });
+        if (!prepared.ok)
+          return { success: false, error: prepared.error.message };
+        if (prepared.value.variantStore) {
+          totalToEmbed = countVariantBacklog(prepared.value);
+          const startedAt = Date.now();
+          if (dryRun)
+            return {
+              success: true,
+              embedded: totalToEmbed,
+              errors: 0,
+              contentionErrors: 0,
+              duration: 0,
+              model: modelUri,
+              searchAvailable: prepared.value.variantStore.searchAvailable,
+              errorSamples: [],
+            };
+          const processed = await embedBacklog({
+            ...prepared.value,
+            onProgress: !options.json
+              ? (embedded, errors) =>
+                  process.stderr.write(
+                    `\rEmbedded ${embedded}/${totalToEmbed} owners; ${errors} errors`
+                  )
+              : undefined,
+          });
+          if (!options.json) process.stderr.write("\n");
+          if (!processed.ok)
+            return { success: false, error: processed.error.message };
+          return {
+            success: true,
+            ...processed.value,
+            contentionErrors: processed.value.contentionErrors ?? 0,
+            duration: (Date.now() - startedAt) / 1000,
+            model: modelUri,
+            searchAvailable: prepared.value.variantStore.searchAvailable,
+            errorSamples: [],
+          };
+        }
+        if (force) {
+          const count = await getActiveChunkCount(db, options.collection);
+          if (!count.ok) return { success: false, error: count.error.message };
+          totalToEmbed = count.value;
+          if (dryRun || !totalToEmbed)
+            return {
+              success: true,
+              embedded: totalToEmbed,
+              errors: 0,
+              contentionErrors: 0,
+              duration: 0,
+              model: modelUri,
+              searchAvailable: vectorIndex.searchAvailable,
+              errorSamples: [],
+            };
+        }
+
         if (!force) {
           const embedFingerprint = getEmbeddingFingerprint({
             modelUri,
@@ -773,7 +809,7 @@ function getActiveChunks(
     const sql = after
       ? `
         SELECT c.mirror_hash as mirrorHash, c.seq, c.text,
-          (SELECT d.title FROM documents d WHERE d.mirror_hash = c.mirror_hash AND d.active = 1 LIMIT 1) as title,
+          (SELECT d.title FROM documents d WHERE d.mirror_hash = c.mirror_hash AND d.active = 1 ORDER BY d.id LIMIT 1) as title,
           'force' as reason
         FROM content_chunks c
         WHERE EXISTS (
@@ -786,7 +822,7 @@ function getActiveChunks(
       `
       : `
         SELECT c.mirror_hash as mirrorHash, c.seq, c.text,
-          (SELECT d.title FROM documents d WHERE d.mirror_hash = c.mirror_hash AND d.active = 1 LIMIT 1) as title,
+          (SELECT d.title FROM documents d WHERE d.mirror_hash = c.mirror_hash AND d.active = 1 ORDER BY d.id LIMIT 1) as title,
           'force' as reason
         FROM content_chunks c
         WHERE EXISTS (

package/src/cli/commands/mcp/install.ts

@@ -6,6 +6,10 @@
 
 import { DEFAULT_INDEX_NAME } from "../../../app/constants.js";
 import { getConfigPaths, toAbsolutePath } from "../../../config/paths.js";
+import {
+  type McpToolProfile,
+  parseMcpToolProfile,
+} from "../../../mcp/tool-profile.js";
 import { CliError } from "../../errors.js";
 import { getGlobals } from "../../program.js";
 import { resolveMcpConfigLocation } from "./config-discovery.js";
@@ -45,6 +49,8 @@ export interface InstallOptions {
   force?: boolean;
   dryRun?: boolean;
   enableWrite?: boolean;
+  /** Advertised tool set written into the registration (`core` or `full`). */
+  toolProfile?: McpToolProfile;
   /** Index identity persisted in the installed MCP command. */
   indexName?: string;
   /** Active GNO config persisted as a canonical absolute path. */
@@ -166,6 +172,18 @@ function safeGetGlobals(): { json: boolean; quiet: boolean } {
   }
 }
 
+/** Validate the requested profile before any config file is read or written. */
+function parseInstallToolProfile(value: unknown): McpToolProfile | undefined {
+  try {
+    return parseMcpToolProfile(value);
+  } catch (error) {
+    throw new CliError(
+      "VALIDATION",
+      error instanceof Error ? error.message : String(error)
+    );
+  }
+}
+
 /**
  * Install gno MCP server.
  */
@@ -175,6 +193,7 @@ export async function installMcp(opts: InstallOptions = {}): Promise<void> {
   const force = opts.force ?? false;
   const dryRun = opts.dryRun ?? false;
   const enableWrite = opts.enableWrite ?? false;
+  const toolProfile = parseInstallToolProfile(opts.toolProfile);
   const indexName = opts.indexName ?? DEFAULT_INDEX_NAME;
   const paths = getConfigPaths();
   const configPath = toAbsolutePath(
@@ -196,6 +215,7 @@ export async function installMcp(opts: InstallOptions = {}): Promise<void> {
   // Build server entry (uses process.execPath, always succeeds)
   const serverEntry = buildMcpServerEntry({
     enableWrite,
+    toolProfile,
     indexName,
     configPath,
     dataDir: toAbsolutePath(paths.dataDir, opts.cwd),

package/src/cli/commands/mcp/paths.ts

@@ -13,6 +13,11 @@ import type { ConnectorWorkspaceEnvironment } from "../../../core/connector-envi
 import { resolveDirs } from "../../../app/constants";
 import { assertValidIndexName } from "../../../app/index-name";
 import { getCurrentGnoEntrypoint } from "../../../core/runtime-entrypoint";
+import {
+  DEFAULT_MCP_TOOL_PROFILE,
+  type McpToolProfile,
+  isMcpToolProfile,
+} from "../../../mcp/tool-profile.js";
 import { getTargetDisplayName } from "./target-display.js";
 
 export { getTargetDisplayName } from "./target-display.js";
@@ -70,6 +75,8 @@ export interface McpServerEntry {
 
 interface McpServerEntryOptions {
   enableWrite?: boolean;
+  /** Advertised MCP tool set the registration starts with; omitted = `full`. */
+  toolProfile?: McpToolProfile;
   indexName?: string;
   configPath?: string;
   dataDir?: string;
@@ -485,7 +492,25 @@ function appendMcpArguments(
     args.push("--config", resolve(options.configPath));
   }
   args.push("mcp");
+  if (options.toolProfile !== undefined) {
+    args.push("--tool-profile", options.toolProfile);
+  }
   if (options.enableWrite) {
     args.push("--enable-write");
   }
 }
+
+/**
+ * The tool profile a registration starts with, read back from its args.
+ * A registration written before the flag existed carries none and runs `full`.
+ */
+export function readMcpToolProfileFromArgs(
+  args: ReadonlyArray<string>
+): McpToolProfile {
+  const index = args.indexOf("--tool-profile");
+  if (index === -1) {
+    return DEFAULT_MCP_TOOL_PROFILE;
+  }
+  const value = args[index + 1];
+  return isMcpToolProfile(value) ? value : DEFAULT_MCP_TOOL_PROFILE;
+}

package/src/cli/commands/mcp/status.ts

@@ -6,6 +6,7 @@
 
 import type { ConnectorWorkspaceEnvironment } from "../../../core/connector-environment";
 import type { McpConnectorVerificationTarget } from "../../../core/connector-verifier";
+import type { McpToolProfile } from "../../../mcp/tool-profile.js";
 import type { StandardMcpEntry } from "./config.js";
 
 import { normalizeConnectorWorkspaceEnvironment } from "../../../core/connector-environment";
@@ -29,6 +30,7 @@ import {
   type McpScope,
   type McpTarget,
   resolveMcpConfigPath,
+  readMcpToolProfileFromArgs,
 } from "./paths.js";
 
 // ─────────────────────────────────────────────────────────────────────────────
@@ -56,6 +58,8 @@ export interface McpTargetStatus {
     args: string[];
     env?: ConnectorWorkspaceEnvironment;
   };
+  /** Tool profile the registration starts with; `full` when its args carry none. */
+  toolProfile?: McpToolProfile;
   error?: string;
 }
 
@@ -234,6 +238,7 @@ export async function checkMcpTargetStatus(
       configPath,
       configured: true,
       serverEntry,
+      toolProfile: readMcpToolProfileFromArgs(serverEntry.args),
     };
     configIdentityByStatus.set(status, configIdentity);
     return status;
@@ -341,6 +346,7 @@ export async function statusMcp(opts: StatusOptions = {}): Promise<void> {
     if (status.configured && status.serverEntry) {
       process.stdout.write(`    Command: ${status.serverEntry.command}\n`);
       process.stdout.write(`    Args: ${status.serverEntry.args.join(" ")}\n`);
+      process.stdout.write(`    Profile: ${status.toolProfile ?? "full"}\n`);
     }
 
     if (status.error) {

package/src/cli/detach.ts

@@ -25,6 +25,7 @@ import type { ResidentStatus } from "../serve/status-model";
 import { VERSION, resolveDirs } from "../app/constants";
 import { toAbsolutePath } from "../config/paths";
 import { atomicWrite } from "../core/file-ops";
+import { RESIDENT_STOP_GRACE_MS } from "../core/shutdown-budget";
 import { isResidentStatus } from "../serve/resident-status";
 import { CliError } from "./errors";
 
@@ -877,7 +878,7 @@ export async function inspectForeignLive(options: {
 export interface StopOptions {
   kind: DetachKind;
   pidFile: string;
-  /** Grace period for SIGTERM before we escalate to SIGKILL. Default 10s. */
+  /** Grace period for SIGTERM before we escalate to SIGKILL. Default 12s. */
   timeoutMs?: number;
   /** Poll interval while waiting for the process to exit. Default 100ms. */
   pollIntervalMs?: number;
@@ -921,7 +922,7 @@ async function waitForExit(
  * unlinks stale pid-files it discovers on entry.
  */
 export async function stopProcess(options: StopOptions): Promise<StopOutcome> {
-  const timeoutMs = options.timeoutMs ?? 10_000;
+  const timeoutMs = options.timeoutMs ?? RESIDENT_STOP_GRACE_MS;
   const pollIntervalMs = options.pollIntervalMs ?? 100;
   const killTimeoutMs = options.killTimeoutMs ?? 2_000;
   const sleep = options.sleep ?? defaultSleep;

package/src/cli/program.ts

@@ -2234,9 +2234,14 @@ function wireMcpCommand(program: Command): void {
       "--enable-write",
       "Enable write operations in installed MCP configuration"
     )
+    .option(
+      "--tool-profile <profile>",
+      "advertised tool set the registration starts with: core (7 read tools + capture/remember with --enable-write) or full (default)"
+    )
     .option("--json", "JSON output")
     .action(async (cmdOpts: Record<string, unknown>) => {
       const target = cmdOpts.target as string;
+      const toolProfile = parseToolProfileOption(cmdOpts.toolProfile);
       const requestedScope = cmdOpts.scope;
 
       // Import MCP_TARGETS for validation
@@ -2278,6 +2283,7 @@ function wireMcpCommand(program: Command): void {
         force: Boolean(cmdOpts.force),
         dryRun: Boolean(cmdOpts.dryRun),
         enableWrite: Boolean(cmdOpts.enableWrite),
+        toolProfile,
         indexName: globals.index,
         configPath: globals.config,
         // Pass undefined if not set, so global --json can take effect

package/src/config/types.ts

@@ -425,9 +425,9 @@ export const ModelConfigSchema = z.object({
   /** Model presets */
   presets: z.array(ModelPresetSchema).default(DEFAULT_MODEL_PRESETS),
   /** Model load timeout in ms */
-  loadTimeout: z.number().default(60_000),
-  /** Inference timeout in ms */
-  inferenceTimeout: z.number().default(30_000),
+  loadTimeout: z.number().int().min(1).max(2_147_483_647).default(60_000),
+  /** Inference timeout in ms, measured from native evaluation start. */
+  inferenceTimeout: z.number().int().min(1).max(2_147_483_647).default(30_000),
   /** Context size used for query expansion generation */
   expandContextSize: z.number().int().min(256).default(2_048),
   /** Keep warm model TTL in ms (5 min) */

package/src/converters/adapters/markitdownTs/adapter.ts

@@ -3,8 +3,6 @@
  * Uses convertBuffer() with bytes for determinism.
  */
 
-import { MarkItDown } from "markitdown-ts";
-
 import type {
   Converter,
   ConvertInput,
@@ -12,6 +10,7 @@ import type {
   ConvertWarning,
 } from "../../types";
 
+import { MarkItDown } from "../../../../vendor/converters/markitdown-ts";
 import {
   adapterError,
   corruptError,

package/src/converters/adapters/officeparser/adapter.ts

@@ -3,8 +3,6 @@
  * Uses parseOffice() v7 API with Buffer for in-memory extraction.
  */
 
-import { parseOffice } from "officeparser";
-
 import type {
   Converter,
   ConvertInput,
@@ -12,6 +10,7 @@ import type {
   ConvertWarning,
 } from "../../types";
 
+import { parseOffice } from "../../../../vendor/converters/officeparser";
 import { adapterError, corruptError, tooLargeError } from "../../errors";
 import { basenameWithoutExt } from "../../path";
 import { ADAPTER_VERSIONS } from "../../versions";

package/src/converters/versions.ts

@@ -2,10 +2,9 @@
  * Centralized converter version tracking.
  *
  * Native converters use our own versioning.
- * Adapter versions MUST match the wrapped npm package version.
- *
- * When updating npm dependencies, update these versions too.
- * Run `bun pm ls markitdown-ts officeparser` to check current versions.
+ * Adapter versions identify upstream code and security-sensitive parser versions.
+ * Update them when changing the vendored distributions or parser dependencies.
+ * Upstream identities live in vendor/converters/upstream-manifest.json.
  */
 
 /** Native converter versions (our own) */
@@ -15,10 +14,9 @@ export const NATIVE_VERSIONS = {
 } as const;
 
 /**
- * Adapter versions - MUST match npm package versions.
- * Update these when running `bun update`.
+ * Include repaired parser versions so existing converted content is invalidated.
  */
 export const ADAPTER_VERSIONS = {
-  "markitdown-ts": "0.0.8",
-  officeparser: "6.0.4",
+  "markitdown-ts": "0.0.10+xlsx.0.20.3",
+  officeparser: "7.8.0+pdfjs.6.3.289",
 } as const;

package/src/core/context-evidence.ts

@@ -1,6 +1,7 @@
 /** Strict indexed-evidence loading for Context Capsule compilation. */
 
 import type { EgressPolicy } from "../config/types";
+import type { RequestHydration } from "../pipeline/hydration";
 import type { SearchResults } from "../pipeline/types";
 import type {
   ActivationIndexSnapshot,
@@ -127,6 +128,7 @@ export interface ContextEvidenceProjectionContext {
 
 export interface ContextEvidenceCompilerDeps<P> {
   store: ContextEvidenceStore;
+  hydration?: Pick<RequestHydration, "getContentBatch" | "getChunksBatch">;
   retrieve: (request: ContextRetrievalRequest) => Promise<SearchResults>;
   projectCanonical: (
     draft: ContextCanonicalPlanDraft<ContextEvidenceValue>,
@@ -329,7 +331,8 @@ const referenceDocument = (
 export const materializeContextEvidenceCandidates = async (
   store: ContextEvidenceStore,
   candidates: ContextRetrievalCandidate[],
-  indexNameInput: string
+  indexNameInput: string,
+  hydration?: ContextEvidenceCompilerDeps<unknown>["hydration"]
 ): Promise<ContextMaterialization<ContextEvidenceValue>[]> => {
   if (candidates.length === 0) return [];
   const indexName = canonicalizeIndexName(indexNameInput);
@@ -363,8 +366,8 @@ export const materializeContextEvidenceCandidates = async (
     ),
   ];
   const [contentResult, chunksResult] = await Promise.all([
-    store.getContentBatch(mirrorHashes),
-    store.getChunksBatch(mirrorHashes),
+    (hydration ?? store).getContentBatch(mirrorHashes),
+    (hydration ?? store).getChunksBatch(mirrorHashes),
   ]);
   const contentByHash = unwrapStore(
     contentResult,
@@ -517,7 +520,8 @@ export const compileContextEvidence = async <P>(
         materializeContextEvidenceCandidates(
           deps.store,
           candidates,
-          input.indexName
+          input.indexName,
+          deps.hydration
         ),
       projectCanonical: (draft) => deps.projectCanonical(draft, fingerprints),
     }