@zenalexa/unicli 0.210.0 → 0.212.1

package/dist/discovery/loader.js

@@ -11,6 +11,25 @@ import { join, extname, basename, dirname } from "node:path";
 import { fileURLToPath, pathToFileURL } from "node:url";
 import yaml from "js-yaml";
 import { registerAdapter } from "../registry.js";
+import { validateAdapterV2 } from "../core/schema-v2.js";
+/**
+ * Upper bound on YAML adapter file size. A legitimate YAML adapter is
+ * under 4 KiB; 256 KiB leaves headroom for generated or commented files
+ * while capping the worst case where a pathological file (billion-laughs
+ * anchor expansion, runaway template, attacker-controlled user dir) could
+ * OOM the loader. Files above this threshold are skipped with a stderr
+ * warning rather than parsed.
+ */
+const MAX_YAML_BYTES = 256 * 1024;
+/**
+ * Environment flag — when set to `warn`, a failed schema-v2 validation
+ * during adapter load emits a stderr warning but keeps loading. Default
+ * `strict` aborts with exit code 78 (CONFIG_ERROR) on any violation —
+ * the hard gate guarantees every registered adapter carries all five
+ * required v2 metadata fields. Set `UNICLI_SCHEMA=warn` to relax during
+ * a migration window.
+ */
+const SCHEMA_MODE = (process.env.UNICLI_SCHEMA ?? "strict").toLowerCase();
 const __dirname = dirname(fileURLToPath(import.meta.url));
 /**
  * Directory resolution — two different concerns:
@@ -79,20 +98,59 @@ export function loadAdaptersFromDir(dir) {
         const commands = {};
         let siteType = "web-api";
         let siteMeta = {};
+        // Load site-level metadata from _site.json if present
+        const siteJsonPath = join(siteDir, "_site.json");
+        if (existsSync(siteJsonPath)) {
+            try {
+                const meta = JSON.parse(readFileSync(siteJsonPath, "utf-8"));
+                if (meta.type)
+                    siteType = meta.type;
+                if (meta.domain)
+                    siteMeta.domain = meta.domain;
+                if (meta.strategy)
+                    siteMeta.strategy = meta.strategy;
+                if (meta.binary)
+                    siteMeta.binary = meta.binary;
+                if (meta.detect)
+                    siteMeta.detect = meta.detect;
+                if (meta.auth_cookies)
+                    siteMeta.authCookies = meta.auth_cookies;
+            }
+            catch {
+                /* ignore malformed _site.json */
+            }
+        }
         for (const file of readdirSync(siteDir)) {
             const ext = extname(file);
             const cmdName = basename(file, ext);
             if (ext === ".yaml" || ext === ".yml") {
                 let parsed;
+                const absPath = join(siteDir, file);
                 try {
-                    const raw = readFileSync(join(siteDir, file), "utf-8");
-                    parsed = yaml.load(raw);
+                    // Enforce a file-size upper bound BEFORE reading into memory so a
+                    // hostile adapter can't OOM the loader through a gigabyte-sized
+                    // YAML. `statSync` is one syscall and avoids touching contents.
+                    const fileSize = statSync(absPath).size;
+                    if (fileSize > MAX_YAML_BYTES) {
+                        console.error(`Warning: Skipping oversized YAML ${absPath} (${fileSize} bytes > ${MAX_YAML_BYTES})`);
+                        continue;
+                    }
+                    const raw = readFileSync(absPath, "utf-8");
+                    // Use CORE_SCHEMA (no JS type tags) + strict schema-style loading.
+                    // js-yaml exposes anchor/alias expansion in all schemas, but
+                    // CORE_SCHEMA blocks `!!js/function`/`!!js/regexp`/`!!js/undefined`
+                    // tags that would let a YAML author execute arbitrary JS on load.
+                    // Anchor-expansion bombs (billion-laughs) are defused by the size
+                    // cap above — the expanded tree can't exceed the input size by
+                    // more than the alias depth, so 256 KiB input → bounded RAM use.
+                    parsed = yaml.load(raw, {
+                        schema: yaml.CORE_SCHEMA,
+                        filename: absPath,
+                    });
                 }
                 catch (err) {
                     const msg = err instanceof Error ? err.message : String(err);
-                    if (process.env.UNICLI_DEBUG) {
-                        console.error(`Warning: Failed to parse ${join(siteDir, file)}: ${msg}`);
-                    }
+                    console.error(`Warning: Failed to parse ${absPath}: ${msg}`);
                     continue;
                 }
                 // detect: field is stored on the adapter manifest for runtime checks,
@@ -120,6 +178,48 @@ export function loadAdaptersFromDir(dir) {
                     siteMeta.passthrough = parsed.passthrough;
                 if (parsed.auth_cookies)
                     siteMeta.authCookies = parsed.auth_cookies;
+                // Skip underscore-prefixed files (internal/metadata, not commands)
+                if (cmdName.startsWith("_"))
+                    continue;
+                // schema-v2 hard gate: every YAML must satisfy the v2 contract
+                // before it's registered. We validate the FULL parsed object (not a
+                // five-field projection) so legacy fields like `pipeline`, `url`,
+                // `params` get type-checked too — a `pipeline:"string"` typo is a
+                // runtime crash and must fail the gate, not be silently carried
+                // through. The defaults for `capabilities`, `minimum_capability`,
+                // `trust`, `confidentiality`, `quarantine` are filled via
+                // {@link migrateToV2} for backward compatibility; if the YAML
+                // already has them, they win.
+                const v2Candidate = {
+                    ...parsed,
+                    name: cmdName,
+                    capabilities: Array.isArray(parsed.capabilities)
+                        ? parsed.capabilities
+                        : [],
+                    minimum_capability: typeof parsed.minimum_capability === "string"
+                        ? parsed.minimum_capability
+                        : "http.fetch",
+                    trust: typeof parsed.trust === "string" ? parsed.trust : "public",
+                    confidentiality: typeof parsed.confidentiality === "string"
+                        ? parsed.confidentiality
+                        : "public",
+                    quarantine: typeof parsed.quarantine === "boolean" ? parsed.quarantine : false,
+                };
+                const v2Result = validateAdapterV2(v2Candidate);
+                if (!v2Result.ok) {
+                    const rel = join(site, file);
+                    const msg = `schema-v2 violation in ${rel}: ${v2Result.error}`;
+                    if (SCHEMA_MODE === "strict") {
+                        console.error(msg);
+                        process.exit(78); // sysexits.h EX_CONFIG
+                    }
+                    // warn mode: ALWAYS write to stderr — the hard-gate claim in
+                    // §1.7 of the FINAL plan (2026-04-14-v212-rethink) is that
+                    // operators see every violation, not just those running with
+                    // UNICLI_DEBUG=1. Silent warnings are what let the "gate is
+                    // theatre" regression slip through the first time.
+                    console.error(`Warning: ${msg}`);
+                }
                 // Parse args from YAML into AdapterArg[]
                 let adapterArgs;
                 if (parsed.args) {
@@ -147,6 +247,8 @@ export function loadAdaptersFromDir(dir) {
                     wait: parsed.wait,
                     extract: parsed.extract,
                     execArgs: parsed.execArgs,
+                    quarantine: parsed.quarantine === true ? true : undefined,
+                    quarantineReason: parsed.quarantineReason,
                 };
                 count++;
             }

package/dist/discovery/loader.js.map

@@ -1 +1 @@
-{"version":3,"file":"loader.js","sourceRoot":"","sources":["../../src/discovery/loader.ts"],"names":[],"mappings":"AAAA;;;;;;;GAOG;AAEH,OAAO,EAAE,YAAY,EAAE,WAAW,EAAE,UAAU,EAAE,QAAQ,EAAE,MAAM,SAAS,CAAC;AAC1E,OAAO,EAAE,IAAI,EAAE,OAAO,EAAE,QAAQ,EAAE,OAAO,EAAE,MAAM,WAAW,CAAC;AAC7D,OAAO,EAAE,aAAa,EAAE,aAAa,EAAE,MAAM,UAAU,CAAC;AACxD,OAAO,IAAI,MAAM,SAAS,CAAC;AAC3B,OAAO,EAAE,eAAe,EAAE,MAAM,gBAAgB,CAAC;AASjD,MAAM,SAAS,GAAG,OAAO,CAAC,aAAa,CAAC,MAAM,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC;AAE1D;;;;;;;;;;;;;;;;;GAiBG;AACH,SAAS,eAAe;IACtB,uDAAuD;IACvD,oEAAoE;IACpE,mEAAmE;IACnE,MAAM,UAAU,GAAG;QACjB,IAAI,CAAC,SAAS,EAAE,IAAI,EAAE,UAAU,CAAC,EAAE,2CAA2C;QAC9E,IAAI,CAAC,SAAS,EAAE,IAAI,EAAE,IAAI,EAAE,KAAK,EAAE,UAAU,CAAC,EAAE,6BAA6B;KAC9E,CAAC;IAEF,qEAAqE;IACrE,IAAI,OAAO,GAAG,UAAU,CAAC,CAAC,CAAC,CAAC;IAC5B,KAAK,MAAM,GAAG,IAAI,UAAU,EAAE,CAAC;QAC7B,IAAI,CAAC,UAAU,CAAC,GAAG,CAAC;YAAE,SAAS;QAC/B,MAAM,OAAO,GAAG,WAAW,CAAC,GAAG,EAAE,EAAE,aAAa,EAAE,IAAI,EAAE,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,EAAE;YACnE,IAAI,CAAC,CAAC,CAAC,WAAW,EAAE,IAAI,CAAC,CAAC,IAAI,CAAC,UAAU,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC,IAAI,CAAC,UAAU,CAAC,GAAG,CAAC;gBACtE,OAAO,KAAK,CAAC;YACf,IAAI,CAAC;gBACH,OAAO,WAAW,CAAC,IAAI,CAAC,GAAG,EAAE,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,IAAI,CACxC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,QAAQ,CAAC,OAAO,CAAC,IAAI,CAAC,CAAC,QAAQ,CAAC,MAAM,CAAC,CACjD,CAAC;YACJ,CAAC;YAAC,MAAM,CAAC;gBACP,OAAO,KAAK,CAAC;YACf,CAAC;QACH,CAAC,CAAC,CAAC;QACH,IAAI,OAAO,EAAE,CAAC;YACZ,OAAO,GAAG,GAAG,CAAC;YACd,MAAM;QACR,CAAC;IACH,CAAC;IAED,2EAA2E;IAC3E,oDAAoD;IACpD,MAAM,KAAK,GAAG,UAAU,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,UAAU,CAAC,CAAC,CAAC,CAAC,IAAI,UAAU,CAAC,CAAC,CAAC,CAAC;IAErE,OAAO,EAAE,OAAO,EAAE,KAAK,EAAE,CAAC;AAC5B,CAAC;AAED,MAAM,EAAE,OAAO,EAAE,gBAAgB,EAAE,KAAK,EAAE,cAAc,EAAE,GAAG,eAAe,EAAE,CAAC;AAC/E,MAAM,QAAQ,GAAG,IAAI,CAAC,OAAO,CAAC,GAAG,CAAC,IAAI,IAAI,GAAG,EAAE,SAAS,EAAE,UAAU,CAAC,CAAC;AAiDtE,yCAAyC;AACzC,MAAM,UAAU,mBAAmB,CAAC,GAAW;IAC7C,IAAI,CAAC,UAAU,CAAC,GAAG,CAAC;QAAE,OAAO,CAAC,CAAC;IAC/B,IAAI,KAAK,GAAG,CAAC,CAAC;IAEd,KAAK,MAAM,IAAI,IAAI,WAAW,CAAC,GAAG,CAAC,EAAE,CAAC;QACpC,IAAI,IAAI,CAAC,UAAU,CAAC,GAAG,CAAC,IAAI,IAAI,CAAC,UAAU,CAAC,GAAG,CAAC;YAAE,SAAS;QAC3D,MAAM,OAAO,GAAG,IAAI,CAAC,GAAG,EAAE,IAAI,CAAC,CAAC;QAChC,IAAI,CAAC,QAAQ,CAAC,OAAO,CAAC,CAAC,WAAW,EAAE;YAAE,SAAS;QAE/C,MAAM,QAAQ,GAAmC,EAAE,CAAC;QACpD,IAAI,QAAQ,GAAgB,SAAwB,CAAC;QACrD,IAAI,QAAQ,GAA6B,EAAE,CAAC;QAE5C,KAAK,MAAM,IAAI,IAAI,WAAW,CAAC,OAAO,CAAC,EAAE,CAAC;YACxC,MAAM,GAAG,GAAG,OAAO,CAAC,IAAI,CAAC,CAAC;YAC1B,MAAM,OAAO,GAAG,QAAQ,CAAC,IAAI,EAAE,GAAG,CAAC,CAAC;YAEpC,IAAI,GAAG,KAAK,OAAO,IAAI,GAAG,KAAK,MAAM,EAAE,CAAC;gBACtC,IAAI,MAAmB,CAAC;gBACxB,IAAI,CAAC;oBACH,MAAM,GAAG,GAAG,YAAY,CAAC,IAAI,CAAC,OAAO,EAAE,IAAI,CAAC,EAAE,OAAO,CAAC,CAAC;oBACvD,MAAM,GAAG,IAAI,CAAC,IAAI,CAAC,GAAG,CAAgB,CAAC;gBACzC,CAAC;gBAAC,OAAO,GAAG,EAAE,CAAC;oBACb,MAAM,GAAG,GAAG,GAAG,YAAY,KAAK,CAAC,CAAC,CAAC,GAAG,CAAC,OAAO,CAAC,CAAC,CAAC,MAAM,CAAC,GAAG,CAAC,CAAC;oBAC7D,IAAI,OAAO,CAAC,GAAG,CAAC,YAAY,EAAE,CAAC;wBAC7B,OAAO,CAAC,KAAK,CACX,4BAA4B,IAAI,CAAC,OAAO,EAAE,IAAI,CAAC,KAAK,GAAG,EAAE,CAC1D,CAAC;oBACJ,CAAC;oBACD,SAAS;gBACX,CAAC;gBAED,sEAAsE;gBACtE,mEAAmE;gBACnE,uEAAuE;gBAEvE,IAAI,MAAM,CAAC,IAAI;oBAAE,QAAQ,GAAG,MAAM,CAAC,IAAmB,CAAC;gBACvD,IAAI,MAAM,CAAC,MAAM;oBAAE,QAAQ,CAAC,MAAM,GAAG,MAAM,CAAC,MAAM,CAAC;gBACnD,IAAI,MAAM,CAAC,QAAQ;oBACjB,QAAQ,CAAC,QAAQ,GAAG,MAAM,CAAC,QAAuC,CAAC;gBACrE,IAAI,MAAM,CAAC,OAAO,KAAK,SAAS;oBAAE,QAAQ,CAAC,OAAO,GAAG,MAAM,CAAC,OAAO,CAAC;gBACpE,IAAI,MAAM,CAAC,MAAM;oBAAE,QAAQ,CAAC,MAAM,GAAG,MAAM,CAAC,MAAM,CAAC;gBACnD,IAAI,MAAM,CAAC,IAAI;oBAAE,QAAQ,CAAC,IAAI,GAAG,MAAM,CAAC,IAAI,CAAC;gBAC7C,IAAI,MAAM,CAAC,MAAM;oBAAE,QAAQ,CAAC,MAAM,GAAG,MAAM,CAAC,MAAM,CAAC;gBACnD,IAAI,MAAM,CAAC,IAAI;oBAAE,QAAQ,CAAC,IAAI,GAAG,MAAM,CAAC,IAA+B,CAAC;gBACxE,IAAI,MAAM,CAAC,WAAW;oBAAE,QAAQ,CAAC,WAAW,GAAG,MAAM,CAAC,WAAW,CAAC;gBAClE,IAAI,MAAM,CAAC,WAAW,KAAK,SAAS;oBAClC,QAAQ,CAAC,WAAW,GAAG,MAAM,CAAC,WAAW,CAAC;gBAC5C,IAAI,MAAM,CAAC,YAAY;oBAAE,QAAQ,CAAC,WAAW,GAAG,MAAM,CAAC,YAAY,CAAC;gBAEpE,yCAAyC;gBACzC,IAAI,WAAqC,CAAC;gBAC1C,IAAI,MAAM,CAAC,IAAI,EAAE,CAAC;oBAChB,WAAW,GAAG,MAAM,CAAC,OAAO,CAAC,MAAM,CAAC,IAAI,CAAC,CAAC,GAAG,CAC3C,CAAC,CAAC,OAAO,EAAE,MAAM,CAAC,EAAE,EAAE,CAAC,CAAC;wBACtB,IAAI,EAAE,OAAO;wBACb,IAAI,EAAG,MAAM,CAAC,IAA2B,IAAI,KAAK;wBAClD,OAAO,EAAE,MAAM,CAAC,OAAO;wBACvB,QAAQ,EAAE,MAAM,CAAC,QAAQ,IAAI,KAAK;wBAClC,UAAU,EAAE,MAAM,CAAC,UAAU,IAAI,KAAK;wBACtC,OAAO,EAAE,MAAM,CAAC,OAAO;wBACvB,WAAW,EAAE,MAAM,CAAC,WAAW;qBAChC,CAAC,CACH,CAAC;gBACJ,CAAC;gBAED,QAAQ,CAAC,OAAO,CAAC,GAAG;oBAClB,IAAI,EAAE,OAAO;oBACb,WAAW,EAAE,MAAM,CAAC,WAAW;oBAC/B,QAAQ,EAAE,MAAM,CAAC,QAAQ;oBACzB,WAAW;oBACX,OAAO,EAAE,MAAM,CAAC,OAAO;oBACvB,MAAM,EAAE,MAAM,CAAC,MAAkC;oBACjD,IAAI,EAAE,MAAM,CAAC,IAAI;oBACjB,GAAG,EAAE,MAAM,CAAC,GAAG;oBACf,MAAM,EAAE,MAAM,CAAC,MAAM;oBACrB,QAAQ,EAAE,MAAM,CAAC,QAAQ;oBACzB,IAAI,EAAE,MAAM,CAAC,IAAI;oBACjB,OAAO,EAAE,MAAM,CAAC,OAAO;oBACvB,QAAQ,EAAE,MAAM,CAAC,QAAQ;iBAC1B,CAAC;gBACF,KAAK,EAAE,CAAC;YACV,CAAC;QACH,CAAC;QAED,IAAI,MAAM,CAAC,IAAI,CAAC,QAAQ,CAAC,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;YACrC,eAAe,CAAC;gBACd,IAAI,EAAE,IAAI;gBACV,IAAI,EAAE,QAAQ;gBACd,QAAQ;gBACR,GAAG,QAAQ;aACZ,CAAC,CAAC;QACL,CAAC;IACH,CAAC;IAED,OAAO,KAAK,CAAC;AACf,CAAC;AAED;;;;;;;;GAQG;AACH,SAAS,cAAc,CAAC,GAAW;IACjC,IAAI,CAAC,UAAU,CAAC,GAAG,CAAC;QAAE,OAAO,EAAE,CAAC;IAChC,MAAM,KAAK,GAAa,EAAE,CAAC;IAC3B,sEAAsE;IACtE,8EAA8E;IAC9E,IAAI,QAAQ,GAAkB,KAAK,CAAC;IACpC,KAAK,MAAM,KAAK,IAAI,WAAW,CAAC,GAAG,CAAC,EAAE,CAAC;QACrC,IAAI,KAAK,CAAC,UAAU,CAAC,GAAG,CAAC,IAAI,KAAK,CAAC,UAAU,CAAC,GAAG,CAAC;YAAE,SAAS;QAC7D,MAAM,QAAQ,GAAG,IAAI,CAAC,GAAG,EAAE,KAAK,CAAC,CAAC;QAClC,IAAI,CAAC,QAAQ,CAAC,QAAQ,CAAC,CAAC,WAAW,EAAE;YAAE,SAAS;QAChD,MAAM,KAAK,GAAG,WAAW,CAAC,QAAQ,CAAC,CAAC,IAAI,CACtC,CAAC,CAAC,EAAE,EAAE,CACJ,CAAC,CAAC,QAAQ,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC,CAAC,QAAQ,CAAC,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC,QAAQ,CAAC,UAAU,CAAC,CACvE,CAAC;QACF,IAAI,KAAK,EAAE,CAAC;YACV,QAAQ,GAAG,KAAK,CAAC;YACjB,MAAM;QACR,CAAC;IACH,CAAC;IAED,KAAK,MAAM,IAAI,IAAI,WAAW,CAAC,GAAG,CAAC,EAAE,CAAC;QACpC,IAAI,IAAI,CAAC,UAAU,CAAC,GAAG,CAAC,IAAI,IAAI,CAAC,UAAU,CAAC,GAAG,CAAC;YAAE,SAAS;QAC3D,MAAM,OAAO,GAAG,IAAI,CAAC,GAAG,EAAE,IAAI,CAAC,CAAC;QAChC,IAAI,CAAC,QAAQ,CAAC,OAAO,CAAC,CAAC,WAAW,EAAE;YAAE,SAAS;QAC/C,KAAK,MAAM,IAAI,IAAI,WAAW,CAAC,OAAO,CAAC,EAAE,CAAC;YACxC,+CAA+C;YAC/C,IAAI,IAAI,CAAC,QAAQ,CAAC,OAAO,CAAC;gBAAE,SAAS;YACrC,IAAI,IAAI,CAAC,QAAQ,CAAC,WAAW,CAAC;gBAAE,SAAS;YACzC,IAAI,IAAI,CAAC,QAAQ,CAAC,SAAS,CAAC;gBAAE,SAAS;YACvC,IAAI,IAAI,CAAC,QAAQ,CAAC,UAAU,CAAC;gBAAE,SAAS;YACxC,IAAI,IAAI,CAAC,QAAQ,CAAC,UAAU,CAAC;gBAAE,SAAS;YACxC,IAAI,OAAO,CAAC,IAAI,CAAC,KAAK,QAAQ,EAAE,CAAC;gBAC/B,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,OAAO,EAAE,IAAI,CAAC,CAAC,CAAC;YAClC,CAAC;QACH,CAAC;IACH,CAAC;IACD,OAAO,KAAK,CAAC;AACf,CAAC;AAED,yEAAyE;AACzE,MAAM,UAAU,eAAe;IAC7B,IAAI,KAAK,GAAG,CAAC,CAAC;IACd,KAAK,IAAI,mBAAmB,CAAC,gBAAgB,CAAC,CAAC;IAC/C,KAAK,IAAI,mBAAmB,CAAC,QAAQ,CAAC,CAAC;IACvC,OAAO,KAAK,CAAC;AACf,CAAC;AAED,uDAAuD;AACvD,MAAM,CAAC,KAAK,UAAU,cAAc;IAClC,MAAM,KAAK,GAAG;QACZ,GAAG,cAAc,CAAC,cAAc,CAAC;QACjC,GAAG,cAAc,CAAC,QAAQ,CAAC;KAC5B,CAAC;IACF,IAAI,KAAK,GAAG,CAAC,CAAC;IACd,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;QACzB,IAAI,CAAC;YACH,MAAM,MAAM,CAAC,aAAa,CAAC,IAAI,CAAC,CAAC,IAAI,CAAC,CAAC;YACvC,KAAK,EAAE,CAAC;QACV,CAAC;QAAC,OAAO,GAAG,EAAE,CAAC;YACb,IAAI,OAAO,CAAC,GAAG,CAAC,YAAY,EAAE,CAAC;gBAC7B,MAAM,GAAG,GAAG,GAAG,YAAY,KAAK,CAAC,CAAC,CAAC,GAAG,CAAC,OAAO,CAAC,CAAC,CAAC,MAAM,CAAC,GAAG,CAAC,CAAC;gBAC7D,OAAO,CAAC,KAAK,CAAC,6BAA6B,IAAI,KAAK,GAAG,EAAE,CAAC,CAAC;YAC7D,CAAC;QACH,CAAC;IACH,CAAC;IACD,OAAO,KAAK,CAAC;AACf,CAAC;AAED;;GAEG;AACH,MAAM,UAAU,cAAc;IAC5B,OAAO,EAAE,OAAO,EAAE,gBAAgB,EAAE,KAAK,EAAE,cAAc,EAAE,CAAC;AAC9D,CAAC"}
+{"version":3,"file":"loader.js","sourceRoot":"","sources":["../../src/discovery/loader.ts"],"names":[],"mappings":"AAAA;;;;;;;GAOG;AAEH,OAAO,EAAE,YAAY,EAAE,WAAW,EAAE,UAAU,EAAE,QAAQ,EAAE,MAAM,SAAS,CAAC;AAC1E,OAAO,EAAE,IAAI,EAAE,OAAO,EAAE,QAAQ,EAAE,OAAO,EAAE,MAAM,WAAW,CAAC;AAC7D,OAAO,EAAE,aAAa,EAAE,aAAa,EAAE,MAAM,UAAU,CAAC;AACxD,OAAO,IAAI,MAAM,SAAS,CAAC;AAC3B,OAAO,EAAE,eAAe,EAAE,MAAM,gBAAgB,CAAC;AACjD,OAAO,EAAE,iBAAiB,EAAE,MAAM,sBAAsB,CAAC;AAEzD;;;;;;;GAOG;AACH,MAAM,cAAc,GAAG,GAAG,GAAG,IAAI,CAAC;AASlC;;;;;;;GAOG;AACH,MAAM,WAAW,GAAG,CAAC,OAAO,CAAC,GAAG,CAAC,aAAa,IAAI,QAAQ,CAAC,CAAC,WAAW,EAAE,CAAC;AAE1E,MAAM,SAAS,GAAG,OAAO,CAAC,aAAa,CAAC,MAAM,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC;AAE1D;;;;;;;;;;;;;;;;;GAiBG;AACH,SAAS,eAAe;IACtB,uDAAuD;IACvD,oEAAoE;IACpE,mEAAmE;IACnE,MAAM,UAAU,GAAG;QACjB,IAAI,CAAC,SAAS,EAAE,IAAI,EAAE,UAAU,CAAC,EAAE,2CAA2C;QAC9E,IAAI,CAAC,SAAS,EAAE,IAAI,EAAE,IAAI,EAAE,KAAK,EAAE,UAAU,CAAC,EAAE,6BAA6B;KAC9E,CAAC;IAEF,qEAAqE;IACrE,IAAI,OAAO,GAAG,UAAU,CAAC,CAAC,CAAC,CAAC;IAC5B,KAAK,MAAM,GAAG,IAAI,UAAU,EAAE,CAAC;QAC7B,IAAI,CAAC,UAAU,CAAC,GAAG,CAAC;YAAE,SAAS;QAC/B,MAAM,OAAO,GAAG,WAAW,CAAC,GAAG,EAAE,EAAE,aAAa,EAAE,IAAI,EAAE,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,EAAE;YACnE,IAAI,CAAC,CAAC,CAAC,WAAW,EAAE,IAAI,CAAC,CAAC,IAAI,CAAC,UAAU,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC,IAAI,CAAC,UAAU,CAAC,GAAG,CAAC;gBACtE,OAAO,KAAK,CAAC;YACf,IAAI,CAAC;gBACH,OAAO,WAAW,CAAC,IAAI,CAAC,GAAG,EAAE,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,IAAI,CACxC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,QAAQ,CAAC,OAAO,CAAC,IAAI,CAAC,CAAC,QAAQ,CAAC,MAAM,CAAC,CACjD,CAAC;YACJ,CAAC;YAAC,MAAM,CAAC;gBACP,OAAO,KAAK,CAAC;YACf,CAAC;QACH,CAAC,CAAC,CAAC;QACH,IAAI,OAAO,EAAE,CAAC;YACZ,OAAO,GAAG,GAAG,CAAC;YACd,MAAM;QACR,CAAC;IACH,CAAC;IAED,2EAA2E;IAC3E,oDAAoD;IACpD,MAAM,KAAK,GAAG,UAAU,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,UAAU,CAAC,CAAC,CAAC,CAAC,IAAI,UAAU,CAAC,CAAC,CAAC,CAAC;IAErE,OAAO,EAAE,OAAO,EAAE,KAAK,EAAE,CAAC;AAC5B,CAAC;AAED,MAAM,EAAE,OAAO,EAAE,gBAAgB,EAAE,KAAK,EAAE,cAAc,EAAE,GAAG,eAAe,EAAE,CAAC;AAC/E,MAAM,QAAQ,GAAG,IAAI,CAAC,OAAO,CAAC,GAAG,CAAC,IAAI,IAAI,GAAG,EAAE,SAAS,EAAE,UAAU,CAAC,CAAC;AAyDtE,yCAAyC;AACzC,MAAM,UAAU,mBAAmB,CAAC,GAAW;IAC7C,IAAI,CAAC,UAAU,CAAC,GAAG,CAAC;QAAE,OAAO,CAAC,CAAC;IAC/B,IAAI,KAAK,GAAG,CAAC,CAAC;IAEd,KAAK,MAAM,IAAI,IAAI,WAAW,CAAC,GAAG,CAAC,EAAE,CAAC;QACpC,IAAI,IAAI,CAAC,UAAU,CAAC,GAAG,CAAC,IAAI,IAAI,CAAC,UAAU,CAAC,GAAG,CAAC;YAAE,SAAS;QAC3D,MAAM,OAAO,GAAG,IAAI,CAAC,GAAG,EAAE,IAAI,CAAC,CAAC;QAChC,IAAI,CAAC,QAAQ,CAAC,OAAO,CAAC,CAAC,WAAW,EAAE;YAAE,SAAS;QAE/C,MAAM,QAAQ,GAAmC,EAAE,CAAC;QACpD,IAAI,QAAQ,GAAgB,SAAwB,CAAC;QACrD,IAAI,QAAQ,GAA6B,EAAE,CAAC;QAE5C,sDAAsD;QACtD,MAAM,YAAY,GAAG,IAAI,CAAC,OAAO,EAAE,YAAY,CAAC,CAAC;QACjD,IAAI,UAAU,CAAC,YAAY,CAAC,EAAE,CAAC;YAC7B,IAAI,CAAC;gBACH,MAAM,IAAI,GAAG,IAAI,CAAC,KAAK,CAAC,YAAY,CAAC,YAAY,EAAE,OAAO,CAAC,CAAC,CAAC;gBAC7D,IAAI,IAAI,CAAC,IAAI;oBAAE,QAAQ,GAAG,IAAI,CAAC,IAAmB,CAAC;gBACnD,IAAI,IAAI,CAAC,MAAM;oBAAE,QAAQ,CAAC,MAAM,GAAG,IAAI,CAAC,MAAM,CAAC;gBAC/C,IAAI,IAAI,CAAC,QAAQ;oBACf,QAAQ,CAAC,QAAQ,GAAG,IAAI,CAAC,QAAuC,CAAC;gBACnE,IAAI,IAAI,CAAC,MAAM;oBAAE,QAAQ,CAAC,MAAM,GAAG,IAAI,CAAC,MAAM,CAAC;gBAC/C,IAAI,IAAI,CAAC,MAAM;oBAAE,QAAQ,CAAC,MAAM,GAAG,IAAI,CAAC,MAAM,CAAC;gBAC/C,IAAI,IAAI,CAAC,YAAY;oBAAE,QAAQ,CAAC,WAAW,GAAG,IAAI,CAAC,YAAY,CAAC;YAClE,CAAC;YAAC,MAAM,CAAC;gBACP,iCAAiC;YACnC,CAAC;QACH,CAAC;QAED,KAAK,MAAM,IAAI,IAAI,WAAW,CAAC,OAAO,CAAC,EAAE,CAAC;YACxC,MAAM,GAAG,GAAG,OAAO,CAAC,IAAI,CAAC,CAAC;YAC1B,MAAM,OAAO,GAAG,QAAQ,CAAC,IAAI,EAAE,GAAG,CAAC,CAAC;YAEpC,IAAI,GAAG,KAAK,OAAO,IAAI,GAAG,KAAK,MAAM,EAAE,CAAC;gBACtC,IAAI,MAAmB,CAAC;gBACxB,MAAM,OAAO,GAAG,IAAI,CAAC,OAAO,EAAE,IAAI,CAAC,CAAC;gBACpC,IAAI,CAAC;oBACH,kEAAkE;oBAClE,gEAAgE;oBAChE,gEAAgE;oBAChE,MAAM,QAAQ,GAAG,QAAQ,CAAC,OAAO,CAAC,CAAC,IAAI,CAAC;oBACxC,IAAI,QAAQ,GAAG,cAAc,EAAE,CAAC;wBAC9B,OAAO,CAAC,KAAK,CACX,oCAAoC,OAAO,KAAK,QAAQ,YAAY,cAAc,GAAG,CACtF,CAAC;wBACF,SAAS;oBACX,CAAC;oBACD,MAAM,GAAG,GAAG,YAAY,CAAC,OAAO,EAAE,OAAO,CAAC,CAAC;oBAC3C,mEAAmE;oBACnE,6DAA6D;oBAC7D,oEAAoE;oBACpE,kEAAkE;oBAClE,kEAAkE;oBAClE,+DAA+D;oBAC/D,iEAAiE;oBACjE,MAAM,GAAG,IAAI,CAAC,IAAI,CAAC,GAAG,EAAE;wBACtB,MAAM,EAAE,IAAI,CAAC,WAAW;wBACxB,QAAQ,EAAE,OAAO;qBAClB,CAAgB,CAAC;gBACpB,CAAC;gBAAC,OAAO,GAAG,EAAE,CAAC;oBACb,MAAM,GAAG,GAAG,GAAG,YAAY,KAAK,CAAC,CAAC,CAAC,GAAG,CAAC,OAAO,CAAC,CAAC,CAAC,MAAM,CAAC,GAAG,CAAC,CAAC;oBAC7D,OAAO,CAAC,KAAK,CAAC,4BAA4B,OAAO,KAAK,GAAG,EAAE,CAAC,CAAC;oBAC7D,SAAS;gBACX,CAAC;gBAED,sEAAsE;gBACtE,mEAAmE;gBACnE,uEAAuE;gBAEvE,IAAI,MAAM,CAAC,IAAI;oBAAE,QAAQ,GAAG,MAAM,CAAC,IAAmB,CAAC;gBACvD,IAAI,MAAM,CAAC,MAAM;oBAAE,QAAQ,CAAC,MAAM,GAAG,MAAM,CAAC,MAAM,CAAC;gBACnD,IAAI,MAAM,CAAC,QAAQ;oBACjB,QAAQ,CAAC,QAAQ,GAAG,MAAM,CAAC,QAAuC,CAAC;gBACrE,IAAI,MAAM,CAAC,OAAO,KAAK,SAAS;oBAAE,QAAQ,CAAC,OAAO,GAAG,MAAM,CAAC,OAAO,CAAC;gBACpE,IAAI,MAAM,CAAC,MAAM;oBAAE,QAAQ,CAAC,MAAM,GAAG,MAAM,CAAC,MAAM,CAAC;gBACnD,IAAI,MAAM,CAAC,IAAI;oBAAE,QAAQ,CAAC,IAAI,GAAG,MAAM,CAAC,IAAI,CAAC;gBAC7C,IAAI,MAAM,CAAC,MAAM;oBAAE,QAAQ,CAAC,MAAM,GAAG,MAAM,CAAC,MAAM,CAAC;gBACnD,IAAI,MAAM,CAAC,IAAI;oBAAE,QAAQ,CAAC,IAAI,GAAG,MAAM,CAAC,IAA+B,CAAC;gBACxE,IAAI,MAAM,CAAC,WAAW;oBAAE,QAAQ,CAAC,WAAW,GAAG,MAAM,CAAC,WAAW,CAAC;gBAClE,IAAI,MAAM,CAAC,WAAW,KAAK,SAAS;oBAClC,QAAQ,CAAC,WAAW,GAAG,MAAM,CAAC,WAAW,CAAC;gBAC5C,IAAI,MAAM,CAAC,YAAY;oBAAE,QAAQ,CAAC,WAAW,GAAG,MAAM,CAAC,YAAY,CAAC;gBAEpE,mEAAmE;gBACnE,IAAI,OAAO,CAAC,UAAU,CAAC,GAAG,CAAC;oBAAE,SAAS;gBAEtC,+DAA+D;gBAC/D,oEAAoE;gBACpE,kEAAkE;gBAClE,kEAAkE;gBAClE,gEAAgE;gBAChE,kEAAkE;gBAClE,0DAA0D;gBAC1D,8DAA8D;gBAC9D,8BAA8B;gBAC9B,MAAM,WAAW,GAA4B;oBAC3C,GAAI,MAA6C;oBACjD,IAAI,EAAE,OAAO;oBACb,YAAY,EAAE,KAAK,CAAC,OAAO,CAAC,MAAM,CAAC,YAAY,CAAC;wBAC9C,CAAC,CAAC,MAAM,CAAC,YAAY;wBACrB,CAAC,CAAC,EAAE;oBACN,kBAAkB,EAChB,OAAO,MAAM,CAAC,kBAAkB,KAAK,QAAQ;wBAC3C,CAAC,CAAC,MAAM,CAAC,kBAAkB;wBAC3B,CAAC,CAAC,YAAY;oBAClB,KAAK,EAAE,OAAO,MAAM,CAAC,KAAK,KAAK,QAAQ,CAAC,CAAC,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC,CAAC,QAAQ;oBACjE,eAAe,EACb,OAAO,MAAM,CAAC,eAAe,KAAK,QAAQ;wBACxC,CAAC,CAAC,MAAM,CAAC,eAAe;wBACxB,CAAC,CAAC,QAAQ;oBACd,UAAU,EACR,OAAO,MAAM,CAAC,UAAU,KAAK,SAAS,CAAC,CAAC,CAAC,MAAM,CAAC,UAAU,CAAC,CAAC,CAAC,KAAK;iBACrE,CAAC;gBACF,MAAM,QAAQ,GAAG,iBAAiB,CAAC,WAAW,CAAC,CAAC;gBAChD,IAAI,CAAC,QAAQ,CAAC,EAAE,EAAE,CAAC;oBACjB,MAAM,GAAG,GAAG,IAAI,CAAC,IAAI,EAAE,IAAI,CAAC,CAAC;oBAC7B,MAAM,GAAG,GAAG,0BAA0B,GAAG,KAAK,QAAQ,CAAC,KAAK,EAAE,CAAC;oBAC/D,IAAI,WAAW,KAAK,QAAQ,EAAE,CAAC;wBAC7B,OAAO,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC;wBACnB,OAAO,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC,CAAC,uBAAuB;oBAC3C,CAAC;oBACD,6DAA6D;oBAC7D,2DAA2D;oBAC3D,6DAA6D;oBAC7D,4DAA4D;oBAC5D,mDAAmD;oBACnD,OAAO,CAAC,KAAK,CAAC,YAAY,GAAG,EAAE,CAAC,CAAC;gBACnC,CAAC;gBAED,yCAAyC;gBACzC,IAAI,WAAqC,CAAC;gBAC1C,IAAI,MAAM,CAAC,IAAI,EAAE,CAAC;oBAChB,WAAW,GAAG,MAAM,CAAC,OAAO,CAAC,MAAM,CAAC,IAAI,CAAC,CAAC,GAAG,CAC3C,CAAC,CAAC,OAAO,EAAE,MAAM,CAAC,EAAE,EAAE,CAAC,CAAC;wBACtB,IAAI,EAAE,OAAO;wBACb,IAAI,EAAG,MAAM,CAAC,IAA2B,IAAI,KAAK;wBAClD,OAAO,EAAE,MAAM,CAAC,OAAO;wBACvB,QAAQ,EAAE,MAAM,CAAC,QAAQ,IAAI,KAAK;wBAClC,UAAU,EAAE,MAAM,CAAC,UAAU,IAAI,KAAK;wBACtC,OAAO,EAAE,MAAM,CAAC,OAAO;wBACvB,WAAW,EAAE,MAAM,CAAC,WAAW;qBAChC,CAAC,CACH,CAAC;gBACJ,CAAC;gBAED,QAAQ,CAAC,OAAO,CAAC,GAAG;oBAClB,IAAI,EAAE,OAAO;oBACb,WAAW,EAAE,MAAM,CAAC,WAAW;oBAC/B,QAAQ,EAAE,MAAM,CAAC,QAAQ;oBACzB,WAAW;oBACX,OAAO,EAAE,MAAM,CAAC,OAAO;oBACvB,MAAM,EAAE,MAAM,CAAC,MAAkC;oBACjD,IAAI,EAAE,MAAM,CAAC,IAAI;oBACjB,GAAG,EAAE,MAAM,CAAC,GAAG;oBACf,MAAM,EAAE,MAAM,CAAC,MAAM;oBACrB,QAAQ,EAAE,MAAM,CAAC,QAAQ;oBACzB,IAAI,EAAE,MAAM,CAAC,IAAI;oBACjB,OAAO,EAAE,MAAM,CAAC,OAAO;oBACvB,QAAQ,EAAE,MAAM,CAAC,QAAQ;oBACzB,UAAU,EAAE,MAAM,CAAC,UAAU,KAAK,IAAI,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,SAAS;oBACzD,gBAAgB,EAAE,MAAM,CAAC,gBAAgB;iBAC1C,CAAC;gBACF,KAAK,EAAE,CAAC;YACV,CAAC;QACH,CAAC;QAED,IAAI,MAAM,CAAC,IAAI,CAAC,QAAQ,CAAC,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;YACrC,eAAe,CAAC;gBACd,IAAI,EAAE,IAAI;gBACV,IAAI,EAAE,QAAQ;gBACd,QAAQ;gBACR,GAAG,QAAQ;aACZ,CAAC,CAAC;QACL,CAAC;IACH,CAAC;IAED,OAAO,KAAK,CAAC;AACf,CAAC;AAED;;;;;;;;GAQG;AACH,SAAS,cAAc,CAAC,GAAW;IACjC,IAAI,CAAC,UAAU,CAAC,GAAG,CAAC;QAAE,OAAO,EAAE,CAAC;IAChC,MAAM,KAAK,GAAa,EAAE,CAAC;IAC3B,sEAAsE;IACtE,8EAA8E;IAC9E,IAAI,QAAQ,GAAkB,KAAK,CAAC;IACpC,KAAK,MAAM,KAAK,IAAI,WAAW,CAAC,GAAG,CAAC,EAAE,CAAC;QACrC,IAAI,KAAK,CAAC,UAAU,CAAC,GAAG,CAAC,IAAI,KAAK,CAAC,UAAU,CAAC,GAAG,CAAC;YAAE,SAAS;QAC7D,MAAM,QAAQ,GAAG,IAAI,CAAC,GAAG,EAAE,KAAK,CAAC,CAAC;QAClC,IAAI,CAAC,QAAQ,CAAC,QAAQ,CAAC,CAAC,WAAW,EAAE;YAAE,SAAS;QAChD,MAAM,KAAK,GAAG,WAAW,CAAC,QAAQ,CAAC,CAAC,IAAI,CACtC,CAAC,CAAC,EAAE,EAAE,CACJ,CAAC,CAAC,QAAQ,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC,CAAC,QAAQ,CAAC,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC,QAAQ,CAAC,UAAU,CAAC,CACvE,CAAC;QACF,IAAI,KAAK,EAAE,CAAC;YACV,QAAQ,GAAG,KAAK,CAAC;YACjB,MAAM;QACR,CAAC;IACH,CAAC;IAED,KAAK,MAAM,IAAI,IAAI,WAAW,CAAC,GAAG,CAAC,EAAE,CAAC;QACpC,IAAI,IAAI,CAAC,UAAU,CAAC,GAAG,CAAC,IAAI,IAAI,CAAC,UAAU,CAAC,GAAG,CAAC;YAAE,SAAS;QAC3D,MAAM,OAAO,GAAG,IAAI,CAAC,GAAG,EAAE,IAAI,CAAC,CAAC;QAChC,IAAI,CAAC,QAAQ,CAAC,OAAO,CAAC,CAAC,WAAW,EAAE;YAAE,SAAS;QAC/C,KAAK,MAAM,IAAI,IAAI,WAAW,CAAC,OAAO,CAAC,EAAE,CAAC;YACxC,+CAA+C;YAC/C,IAAI,IAAI,CAAC,QAAQ,CAAC,OAAO,CAAC;gBAAE,SAAS;YACrC,IAAI,IAAI,CAAC,QAAQ,CAAC,WAAW,CAAC;gBAAE,SAAS;YACzC,IAAI,IAAI,CAAC,QAAQ,CAAC,SAAS,CAAC;gBAAE,SAAS;YACvC,IAAI,IAAI,CAAC,QAAQ,CAAC,UAAU,CAAC;gBAAE,SAAS;YACxC,IAAI,IAAI,CAAC,QAAQ,CAAC,UAAU,CAAC;gBAAE,SAAS;YACxC,IAAI,OAAO,CAAC,IAAI,CAAC,KAAK,QAAQ,EAAE,CAAC;gBAC/B,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,OAAO,EAAE,IAAI,CAAC,CAAC,CAAC;YAClC,CAAC;QACH,CAAC;IACH,CAAC;IACD,OAAO,KAAK,CAAC;AACf,CAAC;AAED,yEAAyE;AACzE,MAAM,UAAU,eAAe;IAC7B,IAAI,KAAK,GAAG,CAAC,CAAC;IACd,KAAK,IAAI,mBAAmB,CAAC,gBAAgB,CAAC,CAAC;IAC/C,KAAK,IAAI,mBAAmB,CAAC,QAAQ,CAAC,CAAC;IACvC,OAAO,KAAK,CAAC;AACf,CAAC;AAED,uDAAuD;AACvD,MAAM,CAAC,KAAK,UAAU,cAAc;IAClC,MAAM,KAAK,GAAG;QACZ,GAAG,cAAc,CAAC,cAAc,CAAC;QACjC,GAAG,cAAc,CAAC,QAAQ,CAAC;KAC5B,CAAC;IACF,IAAI,KAAK,GAAG,CAAC,CAAC;IACd,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;QACzB,IAAI,CAAC;YACH,MAAM,MAAM,CAAC,aAAa,CAAC,IAAI,CAAC,CAAC,IAAI,CAAC,CAAC;YACvC,KAAK,EAAE,CAAC;QACV,CAAC;QAAC,OAAO,GAAG,EAAE,CAAC;YACb,IAAI,OAAO,CAAC,GAAG,CAAC,YAAY,EAAE,CAAC;gBAC7B,MAAM,GAAG,GAAG,GAAG,YAAY,KAAK,CAAC,CAAC,CAAC,GAAG,CAAC,OAAO,CAAC,CAAC,CAAC,MAAM,CAAC,GAAG,CAAC,CAAC;gBAC7D,OAAO,CAAC,KAAK,CAAC,6BAA6B,IAAI,KAAK,GAAG,EAAE,CAAC,CAAC;YAC7D,CAAC;QACH,CAAC;IACH,CAAC;IACD,OAAO,KAAK,CAAC;AACf,CAAC;AAED;;GAEG;AACH,MAAM,UAAU,cAAc;IAC5B,OAAO,EAAE,OAAO,EAAE,gBAAgB,EAAE,KAAK,EAAE,cAAc,EAAE,CAAC;AAC9D,CAAC"}

package/dist/discovery/search.d.ts

@@ -0,0 +1,73 @@
+/**
+ * BM25-based bilingual search engine for command discovery.
+ *
+ * Replaces the naive `String.includes()` filter with a proper information
+ * retrieval algorithm. Designed for ~1000 commands across ~200 sites.
+ *
+ * Architecture:
+ *   1. Build-time: `scripts/build-manifest.js` generates the inverted index
+ *      and IDF values, shipped as `dist/manifest-search.json`.
+ *   2. Runtime: this module loads the index lazily on first search call,
+ *      then scores queries using BM25 with bilingual keyword expansion.
+ *
+ * Performance: <10ms for 1000 documents on a cold index load. The inverted
+ * index is ~50KB — small enough to hold in memory permanently.
+ */
+export interface SearchResult {
+    site: string;
+    command: string;
+    description: string;
+    score: number;
+    usage: string;
+    category: string;
+}
+/** Serialized search index (generated at build time, loaded at runtime). */
+export interface SearchIndex {
+    /** Mapping: term → list of document indices that contain this term */
+    postings: Record<string, number[]>;
+    /** IDF (Inverse Document Frequency) for each term */
+    idf: Record<string, number>;
+    /** All documents with their metadata */
+    documents: Array<{
+        id: string;
+        site: string;
+        command: string;
+        description: string;
+        terms: string[];
+    }>;
+    /** Average document length (term count) across the corpus */
+    avgDl: number;
+    /** Total document count */
+    N: number;
+}
+/**
+ * Build a search index from a manifest object.
+ * Exported for use by the build script.
+ */
+export declare function buildIndex(manifest: {
+    sites: Record<string, {
+        commands: Array<{
+            name: string;
+            description: string;
+        }>;
+    }>;
+}): SearchIndex;
+/**
+ * Search all commands by natural language query.
+ *
+ * Algorithm:
+ *   1. Tokenize query (bilingual-aware)
+ *   2. Expand tokens via alias table
+ *   3. Compute BM25 base score for each candidate document
+ *   4. Apply boost signals (site match, command match, category match)
+ *   5. Return top-K results with usage examples
+ *
+ * @param query - Natural language query (Chinese or English)
+ * @param limit - Maximum results to return (default 5)
+ */
+export declare function search(query: string, limit?: number): SearchResult[];
+/**
+ * Force-reload the search index (useful after index rebuild).
+ */
+export declare function invalidateCache(): void;
+//# sourceMappingURL=search.d.ts.map

package/dist/discovery/search.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"search.d.ts","sourceRoot":"","sources":["../../src/discovery/search.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;GAcG;AAeH,MAAM,WAAW,YAAY;IAC3B,IAAI,EAAE,MAAM,CAAC;IACb,OAAO,EAAE,MAAM,CAAC;IAChB,WAAW,EAAE,MAAM,CAAC;IACpB,KAAK,EAAE,MAAM,CAAC;IACd,KAAK,EAAE,MAAM,CAAC;IACd,QAAQ,EAAE,MAAM,CAAC;CAClB;AAcD,4EAA4E;AAC5E,MAAM,WAAW,WAAW;IAC1B,sEAAsE;IACtE,QAAQ,EAAE,MAAM,CAAC,MAAM,EAAE,MAAM,EAAE,CAAC,CAAC;IACnC,qDAAqD;IACrD,GAAG,EAAE,MAAM,CAAC,MAAM,EAAE,MAAM,CAAC,CAAC;IAC5B,wCAAwC;IACxC,SAAS,EAAE,KAAK,CAAC;QACf,EAAE,EAAE,MAAM,CAAC;QACX,IAAI,EAAE,MAAM,CAAC;QACb,OAAO,EAAE,MAAM,CAAC;QAChB,WAAW,EAAE,MAAM,CAAC;QACpB,KAAK,EAAE,MAAM,EAAE,CAAC;KACjB,CAAC,CAAC;IACH,6DAA6D;IAC7D,KAAK,EAAE,MAAM,CAAC;IACd,2BAA2B;IAC3B,CAAC,EAAE,MAAM,CAAC;CACX;AAuFD;;;GAGG;AACH,wBAAgB,UAAU,CAAC,QAAQ,EAAE;IACnC,KAAK,EAAE,MAAM,CACX,MAAM,EACN;QAAE,QAAQ,EAAE,KAAK,CAAC;YAAE,IAAI,EAAE,MAAM,CAAC;YAAC,WAAW,EAAE,MAAM,CAAA;SAAE,CAAC,CAAA;KAAE,CAC3D,CAAC;CACH,GAAG,WAAW,CAsDd;AA0KD;;;;;;;;;;;;GAYG;AACH,wBAAgB,MAAM,CAAC,KAAK,EAAE,MAAM,EAAE,KAAK,SAAI,GAAG,YAAY,EAAE,CA0G/D;AASD;;GAEG;AACH,wBAAgB,eAAe,IAAI,IAAI,CAEtC"}

package/dist/discovery/search.js

@@ -0,0 +1,392 @@
+/**
+ * BM25-based bilingual search engine for command discovery.
+ *
+ * Replaces the naive `String.includes()` filter with a proper information
+ * retrieval algorithm. Designed for ~1000 commands across ~200 sites.
+ *
+ * Architecture:
+ *   1. Build-time: `scripts/build-manifest.js` generates the inverted index
+ *      and IDF values, shipped as `dist/manifest-search.json`.
+ *   2. Runtime: this module loads the index lazily on first search call,
+ *      then scores queries using BM25 with bilingual keyword expansion.
+ *
+ * Performance: <10ms for 1000 documents on a cold index load. The inverted
+ * index is ~50KB — small enough to hold in memory permanently.
+ */
+import { readFileSync, existsSync } from "node:fs";
+import { join, dirname } from "node:path";
+import { fileURLToPath } from "node:url";
+import { expandToken, tokenizeQuery, SITE_ALIASES, SITE_CATEGORIES, CATEGORY_ALIASES, } from "./aliases.js";
+// ── BM25 Parameters ─────────────────────────────────────────────────────────
+// Standard BM25 tuning. k1 controls term frequency saturation,
+// b controls document length normalization.
+const K1 = 1.2;
+const B = 0.75;
+// ── Hybrid Scoring ──────────────────────────────────────────────────────────
+// StackOne benchmark (Feb 2026, 2700 test cases, 270 tools) found:
+//   Pure BM25 Top-1: 14%
+//   BM25+TF-IDF 20/80 blend Top-1: 21%
+//   Embedding Top-1: 38%
+// We use the 20/80 blend as the base, with domain-specific boosts on top.
+const ALPHA_BM25 = 0.2;
+const ALPHA_TFIDF = 0.8;
+// ── Score Boost Weights ─────────────────────────────────────────────────────
+// Applied on top of the hybrid BM25+TF-IDF base score.
+const BOOST_SITE_EXACT = 15.0; // Query token exactly matches site name
+const BOOST_SITE_ALIAS = 12.0; // Query token's alias matches site name
+const BOOST_CMD_EXACT = 8.0; // Query token exactly matches command name
+const BOOST_CMD_PARTIAL = 3.0; // Query token is substring of command name
+const BOOST_CATEGORY = 2.0; // Query token matches site's category
+// ── Index Management ────────────────────────────────────────────────────────
+let cachedIndex = null;
+/**
+ * Resolve the path to the pre-built search index.
+ * Falls back to building one on-the-fly from the manifest if the
+ * pre-built index doesn't exist.
+ */
+function getIndexPath() {
+    const __dirname = dirname(fileURLToPath(import.meta.url));
+    return join(__dirname, "..", "..", "dist", "manifest-search.json");
+}
+/**
+ * Load or build the search index. Called lazily on first search.
+ */
+function loadIndex() {
+    if (cachedIndex)
+        return cachedIndex;
+    const indexPath = getIndexPath();
+    if (existsSync(indexPath)) {
+        cachedIndex = JSON.parse(readFileSync(indexPath, "utf-8"));
+        return cachedIndex;
+    }
+    // Fallback: build index on-the-fly from manifest.json
+    cachedIndex = buildIndexFromManifest();
+    return cachedIndex;
+}
+/**
+ * Build a search index from the manifest.json file.
+ * Used when the pre-built search index doesn't exist (dev mode).
+ */
+function buildIndexFromManifest() {
+    const __dirname = dirname(fileURLToPath(import.meta.url));
+    const manifestPath = join(__dirname, "..", "..", "dist", "manifest.json");
+    if (!existsSync(manifestPath)) {
+        // No manifest either — search will silently return zero results.
+        // Emit an actionable hint to stderr so CI/dev failures point to the fix.
+        process.stderr.write("[unicli search] Missing dist/manifest-search.json and dist/manifest.json. " +
+            "Run: npm run build:manifest\n");
+        return { postings: {}, idf: {}, documents: [], avgDl: 0, N: 0 };
+    }
+    const manifest = JSON.parse(readFileSync(manifestPath, "utf-8"));
+    return buildIndex(manifest);
+}
+/**
+ * Build a search index from a manifest object.
+ * Exported for use by the build script.
+ */
+export function buildIndex(manifest) {
+    const documents = [];
+    for (const [site, info] of Object.entries(manifest.sites)) {
+        for (const cmd of info.commands) {
+            const terms = tokenizeDocument(site, cmd.name, cmd.description ?? "");
+            documents.push({
+                id: `${site}/${cmd.name}`,
+                site,
+                command: cmd.name,
+                description: cmd.description ?? "",
+                terms,
+                termCount: terms.length,
+            });
+        }
+    }
+    const N = documents.length;
+    const avgDl = N > 0 ? documents.reduce((sum, d) => sum + d.termCount, 0) / N : 0;
+    // Build inverted index
+    const postings = {};
+    for (let i = 0; i < documents.length; i++) {
+        const seen = new Set();
+        for (const term of documents[i].terms) {
+            if (seen.has(term))
+                continue;
+            seen.add(term);
+            if (!postings[term])
+                postings[term] = [];
+            postings[term].push(i);
+        }
+    }
+    // Compute IDF for each term
+    const idf = {};
+    for (const [term, docs] of Object.entries(postings)) {
+        // BM25 IDF: log((N - df + 0.5) / (df + 0.5) + 1)
+        const df = docs.length;
+        idf[term] = Math.log((N - df + 0.5) / (df + 0.5) + 1);
+    }
+    return {
+        postings,
+        idf,
+        documents: documents.map((d) => ({
+            id: d.id,
+            site: d.site,
+            command: d.command,
+            description: d.description,
+            terms: d.terms,
+        })),
+        avgDl,
+        N,
+    };
+}
+// Minimal English stopwords — same set used in query tokenization
+const DOC_STOPWORDS = new Set([
+    "the",
+    "a",
+    "an",
+    "of",
+    "for",
+    "and",
+    "or",
+    "in",
+    "to",
+    "on",
+    "by",
+    "is",
+    "it",
+    "be",
+    "as",
+    "at",
+    "so",
+    "we",
+    "he",
+    "do",
+    "no",
+    "if",
+    "up",
+    "my",
+]);
+// Regex to keep alphanumeric, CJK (all planes), and whitespace.
+// Uses the `u` flag for supplementary plane support.
+const DOC_CLEAN_REGEX = /[^a-z0-9\u4e00-\u9fff\u3400-\u4dbf\uf900-\ufaff\u{20000}-\u{2a6df}\u{2a700}-\u{2b73f}\u{2b740}-\u{2b81f}\u{2b820}-\u{2ceaf}\u{2ceb0}-\u{2ebef}\u{30000}-\u{3134f}\u{31350}-\u{323af}\s]/gu;
+/**
+ * Tokenize a document (site + command + description) into search terms.
+ * Produces lowercase terms, splits on hyphens and underscores.
+ * Applies NFKC normalization and stopword filtering for alignment with
+ * the query tokenizer.
+ */
+function tokenizeDocument(site, command, description) {
+    const terms = [];
+    // Site name and its parts
+    const siteParts = site.toLowerCase().split(/[-_]/);
+    terms.push(site.toLowerCase(), ...siteParts);
+    // Command name and its parts
+    const cmdParts = command.toLowerCase().split(/[-_]/);
+    terms.push(command.toLowerCase(), ...cmdParts);
+    // NFKC normalize description (full-width → half-width, etc.)
+    const normalizedDesc = description.normalize("NFKC");
+    // Description words (lowercase, filter short words and stopwords)
+    const descWords = normalizedDesc
+        .toLowerCase()
+        .replace(DOC_CLEAN_REGEX, " ")
+        .split(/\s+/)
+        .filter((w) => w.length > 1 && !DOC_STOPWORDS.has(w));
+    terms.push(...descWords);
+    // Category as a term
+    const category = SITE_CATEGORIES.get(site);
+    if (category)
+        terms.push(category);
+    return terms;
+}
+// ── BM25 Scoring ────────────────────────────────────────────────────────────
+/**
+ * Compute BM25 score for a single document given query terms.
+ */
+function bm25Score(docTerms, docLength, queryTerms, index) {
+    let score = 0;
+    // Count term frequencies in this document
+    const tf = new Map();
+    for (const term of docTerms) {
+        tf.set(term, (tf.get(term) ?? 0) + 1);
+    }
+    for (const qt of queryTerms) {
+        const termIdf = index.idf[qt];
+        if (termIdf === undefined)
+            continue; // term not in corpus
+        const termTf = tf.get(qt) ?? 0;
+        if (termTf === 0)
+            continue;
+        // BM25 TF component: (tf * (k1 + 1)) / (tf + k1 * (1 - b + b * dl/avgdl))
+        const numerator = termTf * (K1 + 1);
+        const denominator = termTf + K1 * (1 - B + B * (docLength / index.avgDl));
+        score += termIdf * (numerator / denominator);
+    }
+    return score;
+}
+// ── TF-IDF Cosine Similarity ────────────────────────────────────────────────
+/**
+ * Compute TF-IDF cosine similarity between a query and a document.
+ *
+ * TF-IDF for a term t in document d:
+ *   tf(t,d) = count(t in d) / |d|
+ *   tfidf(t,d) = tf(t,d) * idf(t)
+ *
+ * Cosine similarity = dot(query_vec, doc_vec) / (|query_vec| * |doc_vec|)
+ */
+function tfidfCosine(docTerms, queryTerms, index) {
+    const docLen = docTerms.length;
+    if (docLen === 0)
+        return 0;
+    // Build full doc TF map
+    const docTf = new Map();
+    for (const term of docTerms) {
+        docTf.set(term, (docTf.get(term) ?? 0) + 1);
+    }
+    // Compute full document norm (all terms, not just query overlap)
+    let docNormSq = 0;
+    for (const [term, count] of docTf) {
+        const idfVal = index.idf[term];
+        if (idfVal === undefined)
+            continue;
+        const w = (count / docLen) * idfVal;
+        docNormSq += w * w;
+    }
+    // Compute query norm and dot product
+    let dotProduct = 0;
+    let queryNormSq = 0;
+    for (const qt of queryTerms) {
+        const idfVal = index.idf[qt];
+        if (idfVal === undefined)
+            continue;
+        // Query TF-IDF: binary tf (1) × idf
+        const queryWeight = idfVal;
+        queryNormSq += queryWeight * queryWeight;
+        // Doc TF-IDF: normalized tf × idf
+        const rawTf = docTf.get(qt) ?? 0;
+        if (rawTf === 0)
+            continue;
+        const docWeight = (rawTf / docLen) * idfVal;
+        dotProduct += queryWeight * docWeight;
+    }
+    const normProduct = Math.sqrt(queryNormSq) * Math.sqrt(docNormSq);
+    if (normProduct === 0)
+        return 0;
+    return dotProduct / normProduct;
+}
+// ── Main Search Function ────────────────────────────────────────────────────
+/**
+ * Search all commands by natural language query.
+ *
+ * Algorithm:
+ *   1. Tokenize query (bilingual-aware)
+ *   2. Expand tokens via alias table
+ *   3. Compute BM25 base score for each candidate document
+ *   4. Apply boost signals (site match, command match, category match)
+ *   5. Return top-K results with usage examples
+ *
+ * @param query - Natural language query (Chinese or English)
+ * @param limit - Maximum results to return (default 5)
+ */
+export function search(query, limit = 5) {
+    const index = loadIndex();
+    if (index.N === 0)
+        return [];
+    // Step 1: Tokenize
+    const rawTokens = tokenizeQuery(query);
+    // Step 2: Expand via aliases
+    const expandedTerms = [];
+    const siteHints = []; // Directly matched site names
+    const categoryHints = []; // Matched categories
+    for (const token of rawTokens) {
+        const expanded = expandToken(token);
+        expandedTerms.push(...expanded);
+        // Check if this token resolves to a site name
+        const siteMatch = SITE_ALIASES.get(token) ?? SITE_ALIASES.get(token.toLowerCase());
+        if (siteMatch)
+            siteHints.push(siteMatch);
+        // Check if this token is directly a known site
+        if (index.documents.some((d) => d.site === token.toLowerCase())) {
+            siteHints.push(token.toLowerCase());
+        }
+        // Check category alias
+        const catMatch = CATEGORY_ALIASES.get(token) ?? CATEGORY_ALIASES.get(token.toLowerCase());
+        if (catMatch)
+            categoryHints.push(catMatch);
+    }
+    const queryTerms = [...new Set(expandedTerms.map((t) => t.toLowerCase()))];
+    // Step 3: Find candidate documents (union of posting lists)
+    const candidateSet = new Set();
+    for (const qt of queryTerms) {
+        const postings = index.postings[qt];
+        if (postings) {
+            for (const docIdx of postings) {
+                candidateSet.add(docIdx);
+            }
+        }
+    }
+    // If site hints exist, also add ALL commands for those sites
+    if (siteHints.length > 0) {
+        for (let i = 0; i < index.documents.length; i++) {
+            if (siteHints.includes(index.documents[i].site)) {
+                candidateSet.add(i);
+            }
+        }
+    }
+    if (candidateSet.size === 0)
+        return [];
+    // Step 4: Score candidates using hybrid BM25 + TF-IDF
+    const scored = [];
+    for (const idx of candidateSet) {
+        const doc = index.documents[idx];
+        // Hybrid base: alpha-blend BM25 and TF-IDF cosine similarity.
+        // BM25 scores are unbounded; cosine is [0,1]. We scale cosine by the
+        // average BM25 score across candidates to keep the blend balanced.
+        const bm25 = bm25Score(doc.terms, doc.terms.length, queryTerms, index);
+        const tfidf = tfidfCosine(doc.terms, queryTerms, index);
+        let score = ALPHA_BM25 * bm25 + ALPHA_TFIDF * tfidf * 10;
+        // Boost: exact site name match
+        if (siteHints.includes(doc.site)) {
+            score += BOOST_SITE_EXACT;
+        }
+        // Boost: alias-resolved site match
+        for (const qt of queryTerms) {
+            if (qt === doc.site)
+                score += BOOST_SITE_ALIAS;
+            if (qt === doc.command)
+                score += BOOST_CMD_EXACT;
+            if (doc.command.includes(qt) && qt.length > 2)
+                score += BOOST_CMD_PARTIAL;
+        }
+        // Boost: category match
+        const docCategory = SITE_CATEGORIES.get(doc.site);
+        if (docCategory && categoryHints.includes(docCategory)) {
+            score += BOOST_CATEGORY;
+        }
+        if (score > 0)
+            scored.push({ idx, score });
+    }
+    // Step 5: Sort and return top-K
+    scored.sort((a, b) => b.score - a.score);
+    const topK = scored.slice(0, limit);
+    return topK.map(({ idx, score }) => {
+        const doc = index.documents[idx];
+        const category = SITE_CATEGORIES.get(doc.site) ?? "other";
+        return {
+            site: doc.site,
+            command: doc.command,
+            description: doc.description,
+            score: Math.round(score * 100) / 100,
+            usage: buildUsageExample(doc.site, doc.command),
+            category,
+        };
+    });
+}
+/**
+ * Build a usage example string for a command.
+ */
+function buildUsageExample(site, command) {
+    return `unicli ${site} ${command}`;
+}
+/**
+ * Force-reload the search index (useful after index rebuild).
+ */
+export function invalidateCache() {
+    cachedIndex = null;
+}
+//# sourceMappingURL=search.js.map