modelfunction-cli 0.3.0 → 0.4.0

package/README.md

@@ -64,7 +64,7 @@ INTERNAL_LLM_KEY=... mf eval --suite mf.suite.json \
 
 | 命令 | 作用 |
 | --- | --- |
-| `mf scan` | 扫描工具定义（function.json / OpenAPI / MCP 配置 / Python `@tool`）；`--probe` 连 MCP server；`--from` 导入运行时工具数组；草稿按来源分组（`profiles`）并列出未覆盖的 skill |
+| `mf scan` | 扫描工具定义（function.json / OpenAPI / MCP 配置 / Python `@tool` / JS·TS 声明与 TypeBox·zod schema）；`--probe` 连 MCP server（运行时结果覆盖静态推断）；`--from` 导入运行时工具数组；草稿按来源分组（`profiles`）并列出未覆盖的 skill |
 | `mf eval` | 跑评分卡；`--repeat` 多次取多数并报不稳定用例；`--fail-under` 阈值；`--baseline` + `--fail-on-regression` 回归门禁；`--profile` 按挂载组分别评分 |
 | `mf case add` | 写入用例（默认 `source=agent`、`reviewed=false`） |
 | `mf coverage` | 每工具用例数、参数覆盖、负例比例、疑似混淆对；有 `profiles` 时按组统计 |

package/lib/js-tools.js

@@ -0,0 +1,954 @@
+// Static extraction of tool definitions from JS/TS source. This never executes
+// the file: it parses the subset of expressions that real registries use —
+// object/array literals, `const` references to those literals, and the TypeBox
+// (`Type.Object`) / zod (`z.object`) builders. Anything else is reported as
+// unresolved instead of being guessed, because a wrong schema is worse than a
+// missing tool.
+const MAX_VALUE_DEPTH = 24;
+const MAX_RESOLVE_DEPTH = 8;
+const MAX_ARRAY_ATTEMPTS = 300;
+
+const PREFILTER = /inputSchema|parameters\s*:|registerTool|\.tool\s*\(|buildTool|defineTool|createTool|Type\.Object|[tz]\.object/;
+const TOOL_BUILDERS = new Set(["buildTool", "defineTool", "createTool", "makeTool"]);
+// `async () => {}` and `function () {}` otherwise parse as a call named
+// "async"/"function" and swallow the rest of the argument list.
+const RESERVED = new Set([
+  "async", "await", "function", "new", "return", "if", "else", "for", "while",
+  "do", "switch", "case", "try", "catch", "finally", "throw", "typeof", "void",
+  "delete", "in", "of", "instanceof", "class", "const", "let", "var", "export",
+  "import", "yield", "this", "super"
+]);
+const SCHEMA_KEYS = ["inputSchema", "parameters", "input", "parametersSchema", "schema"];
+
+export function extractJsTools(src, { externals, onUnresolvedName } = {}) {
+  if (typeof src !== "string" || !PREFILTER.test(src)) return [];
+  const ctx = { src, bindings: collectBindings(src), parsed: new Map(), externals, onUnresolvedName };
+  const tools = [];
+  const seen = new Set();
+  const push = (tool) => {
+    if (!tool || !tool.name || seen.has(tool.name)) return;
+    seen.add(tool.name);
+    tools.push(tool);
+  };
+  let attempts = 0;
+
+  for (const match of src.matchAll(/([A-Za-z_$][\w$]*)\.(tool|registerTool)\s*\(/g)) {
+    const parsed = parseArgs(src, match.index + match[0].length - 1);
+    if (parsed) push(toolFromArgs(parsed.items, ctx));
+  }
+  for (const match of src.matchAll(/(?<![.\w$])registerTool\s*\(/g)) {
+    const parsed = parseArgs(src, match.index + match[0].length - 1);
+    if (parsed) push(toolFromArgs(parsed.items, ctx));
+  }
+  for (const match of src.matchAll(/(?<![.\w$])([A-Za-z_$][\w$]*)\s*\(/g)) {
+    if (!TOOL_BUILDERS.has(match[1])) continue;
+    const parsed = parseArgs(src, match.index + match[0].length - 1);
+    if (!parsed || !parsed.items[0] || parsed.items[0].t !== "obj") continue;
+    push(toolFromObject(parsed.items[0], ctx));
+  }
+  for (const match of src.matchAll(/(?:export\s+)?(?:const|let|var)\s+[A-Za-z_$][\w$]*\s*(?::[^=\n]{0,200})?=\s*\{/g)) {
+    const parsed = parseValue(src, match.index + match[0].length - 1, 0);
+    if (parsed && parsed.node.t === "obj") push(toolFromObject(parsed.node, ctx));
+  }
+  for (const match of src.matchAll(/\[\s*\{/g)) {
+    if (attempts >= MAX_ARRAY_ATTEMPTS) break;
+    attempts += 1;
+    const parsed = parseValue(src, match.index, 0);
+    if (!parsed || parsed.node.t !== "arr") continue;
+    for (const item of parsed.node.items) {
+      if (item && item.t === "obj") push(toolFromObject(item, ctx));
+    }
+  }
+  return tools;
+}
+
+function toolFromArgs(args, ctx) {
+  const name = stringOf(args[0], ctx);
+  if (!name) return null;
+  let description = "";
+  let parameters = null;
+  for (const arg of args.slice(1)) {
+    if (!arg) continue;
+    if (arg.t === "str") {
+      if (!description) description = arg.value;
+      continue;
+    }
+    if (arg.t !== "obj") continue;
+    const fromObject = toolFromObject(arg, ctx, { name, description });
+    if (fromObject) return fromObject;
+  }
+  for (const arg of args.slice(1)) {
+    if (!arg) continue;
+    const { schema, resolved } = schemaOf(arg, ctx);
+    if (!resolved || !isSchema(schema)) continue;
+    parameters = schema;
+    break;
+  }
+  if (!parameters) return null;
+  return { name, description, parameters };
+}
+
+function toolFromObject(node, ctx, preset = {}) {
+  const name = preset.name ?? stringOf(node.props.get("name"), ctx) ?? stringOf(node.props.get("toolName"), ctx);
+  if (!name) return undefined;
+  const description =
+    stringOf(node.props.get("description"), ctx) ??
+    stringOf(node.props.get("summary"), ctx) ??
+    preset.description ??
+    "";
+  let schemaNode = null;
+  for (const key of SCHEMA_KEYS) {
+    if (node.props.has(key)) {
+      schemaNode = node.props.get(key);
+      break;
+    }
+  }
+  if (!schemaNode) return undefined;
+  const { schema, resolved } = schemaOf(schemaNode, ctx);
+  if (!resolved || !isSchema(schema)) return undefined;
+  // An empty schema with no description and no tool behaviour is a config
+  // object that happens to be shaped like a tool (`{ name: "mcp" }`).
+  const toolish =
+    node.props.has("call") ||
+    node.props.has("prompt") ||
+    node.props.has("userFacingName") ||
+    node.props.has("isEnabled") ||
+    node.props.has("searchHint");
+  if (!description && Object.keys(schema.properties ?? {}).length === 0 && !toolish) return undefined;
+  return { name, description, parameters: schema };
+}
+
+function isSchema(value) {
+  return Boolean(value) && typeof value === "object" && !Array.isArray(value) && Object.keys(value).length > 0;
+}
+
+function stringOf(node, ctx) {
+  if (!node) return null;
+  if (node.t === "str") return node.value;
+  if (node.t === "ident") {
+    // `name: CONFIG_TOOL_NAME` — a local const, or a repo-wide uppercase
+    // constant indexed by the scanner across files.
+    const local = resolveIdent(node.name, ctx);
+    if (local && local.t === "str") return local.value;
+    const external = ctx?.externals?.get(node.name);
+    if (typeof external === "string") return external;
+    ctx?.onUnresolvedName?.(node.name);
+    return null;
+  }
+  if (node.t === "obj") {
+    const value = node.props.get("value");
+    if (value && value.t === "str") return value.value;
+  }
+  return null;
+}
+
+// ---- schema construction -------------------------------------------------
+
+function schemaOf(node, ctx, depth = 0) {
+  if (!node || depth > MAX_VALUE_DEPTH) return { schema: {}, resolved: false };
+  if (node.t === "ident") {
+    const resolvedNode = resolveIdent(node.name, ctx, depth);
+    if (!resolvedNode) return { schema: {}, resolved: false };
+    return schemaOf(resolvedNode, ctx, depth + 1);
+  }
+  if (node.t === "obj") {
+    if (node.props.has("type") || node.props.has("properties") || node.props.has("items")) {
+      return { schema: jsonSchemaLiteral(node, ctx, depth), resolved: true };
+    }
+    // MCP SDK shape argument: server.tool("x", "desc", { query: z.string() }).
+    const rawShape = rawShapeSchema(node, ctx, depth);
+    return rawShape ? { schema: rawShape, resolved: true } : { schema: {}, resolved: false };
+  }
+  if (node.t !== "call") return { schema: {}, resolved: false };
+
+  const args = node.args ?? [];
+  const callee = node.callee;
+  let schema = null;
+  let optional = false;
+
+  switch (callee) {
+    case "Type.Object":
+    case "t.Object":
+    case "z.object":
+    case "z.strictObject":
+    case "z.looseObject":
+      schema = objectSchema(args[0], ctx, depth);
+      break;
+    case "Type.String":
+    case "t.String":
+    case "z.string":
+      schema = { type: "string" };
+      break;
+    case "Type.Number":
+    case "t.Number":
+    case "z.number":
+      schema = { type: "number" };
+      break;
+    case "Type.Integer":
+    case "t.Integer":
+    case "z.int":
+      schema = { type: "integer" };
+      break;
+    case "z.coerce.number":
+      schema = { type: "number" };
+      break;
+    case "z.coerce.string":
+      schema = { type: "string" };
+      break;
+    case "z.coerce.boolean":
+      schema = { type: "boolean" };
+      break;
+    case "Type.Boolean":
+    case "t.Boolean":
+    case "z.boolean":
+      schema = { type: "boolean" };
+      break;
+    case "Type.Array":
+    case "t.Array":
+    case "z.array": {
+      const item = schemaOf(args[0], ctx, depth + 1);
+      schema = { type: "array", items: item.schema };
+      break;
+    }
+    case "Type.Optional":
+    case "z.optional": {
+      const inner = schemaOf(args[0], ctx, depth + 1);
+      return { ...inner, optional: true };
+    }
+    case "Type.Literal":
+    case "t.Literal":
+    case "z.literal": {
+      const value = plainOf(args[0]);
+      schema = value === undefined ? {} : { enum: [value] };
+      break;
+    }
+    case "z.enum": {
+      const values = plainOf(args[0]);
+      schema = Array.isArray(values) ? { type: "string", enum: values } : {};
+      break;
+    }
+    case "Type.Union":
+    case "t.Union":
+    case "z.union": {
+      const nodes = args.length === 1 && args[0] && args[0].t === "arr" ? args[0].items : args;
+      const literals = nodes.map((item) => plainOf(item));
+      if (literals.length > 0 && literals.every((item) => item !== undefined)) {
+        schema = { enum: literals };
+        break;
+      }
+      // z.union([z.string(), z.boolean()]) — keep the types we can name.
+      const types = [
+        ...new Set(nodes.map((item) => schemaOf(item, ctx, depth + 1).schema?.type).filter((t) => typeof t === "string"))
+      ];
+      schema = types.length > 0 ? { type: types.length === 1 ? types[0] : types } : {};
+      break;
+    }
+    case "lazySchema":
+    case "memoize":
+    case "lazy":
+    case "once":
+      return schemaOf(args[0], ctx, depth + 1);
+    case "Type.Record":
+    case "t.Record":
+    case "z.record":
+      schema = { type: "object", additionalProperties: true };
+      break;
+    case "Type.Any":
+    case "Type.Unknown":
+    case "z.any":
+    case "z.unknown":
+      schema = {};
+      break;
+    default: {
+      // Factory call: `inputSchema()` where
+      // `inputSchema = lazySchema(() => z.object({...}))`.
+      if (/^[A-Za-z_$][\w$]*$/.test(callee)) {
+        const target = resolveIdent(callee, ctx, depth);
+        if (target) return schemaOf(target, ctx, depth + 1);
+      }
+      // Wrapper helpers take exactly one schema and add semantics to it
+      // (`semanticNumber(z.number().int())`). Unwrap only when exactly one
+      // argument resolves to a schema, so arrays of schemas are left alone.
+      const inner = args
+        .filter(Boolean)
+        .map((arg) => schemaOf(arg, ctx, depth + 1))
+        .filter((candidate) => candidate.resolved && isSchema(candidate.schema));
+      if (inner.length === 1) return inner[0];
+      return { schema: {}, resolved: false };
+    }
+  }
+
+  // TypeBox carries options as the first object argument, e.g.
+  // Type.String({ description: "File path", format: "uuid" }).
+  if (schema && args[0] && args[0].t === "obj" && /^(Type|t)\./.test(callee)) {
+    applyOptions(schema, args[0]);
+  }
+  for (const chained of node.calls ?? []) {
+    if (applyMethod(schema, chained) === "optional") optional = true;
+  }
+  return { schema, resolved: true, optional };
+}
+
+function rawShapeSchema(node, ctx, depth) {
+  const properties = {};
+  const required = [];
+  for (const [key, valueNode] of node.props) {
+    if (!valueNode || !["call", "ident", "obj"].includes(valueNode.t)) return null;
+    const { schema, resolved, optional } = schemaOf(valueNode, ctx, depth + 1);
+    if (!resolved) return null;
+    properties[key] = schema;
+    if (!optional) required.push(key);
+  }
+  if (Object.keys(properties).length === 0) return null;
+  return { type: "object", properties, required };
+}
+
+function objectSchema(node, ctx, depth) {
+  if (!node) return { type: "object", properties: {} };
+  const resolved = node.t === "ident" ? resolveIdent(node.name, ctx, depth) : node;
+  if (!resolved || resolved.t !== "obj") return { type: "object", properties: {} };
+  const properties = {};
+  const required = [];
+  for (const [key, valueNode] of resolved.props) {
+    const { schema, optional } = schemaOf(valueNode, ctx, depth + 1);
+    properties[key] = schema;
+    if (!optional) required.push(key);
+  }
+  return { type: "object", properties, required };
+}
+
+function jsonSchemaLiteral(node, ctx, depth) {
+  const schema = {};
+  const type = plainOf(node.props.get("type"));
+  if (typeof type === "string") schema.type = type;
+  const description = plainOf(node.props.get("description"));
+  if (typeof description === "string") schema.description = description;
+  const format = plainOf(node.props.get("format"));
+  if (typeof format === "string") schema.format = format;
+  const enumValues = plainOf(node.props.get("enum"));
+  if (Array.isArray(enumValues)) schema.enum = enumValues;
+  for (const key of ["default", "minimum", "maximum", "minLength", "maxLength", "pattern"]) {
+    const value = plainOf(node.props.get(key));
+    if (value !== undefined) schema[key] = value;
+  }
+  if (node.props.has("items")) {
+    const item = schemaOf(node.props.get("items"), ctx, depth + 1);
+    schema.items = item.schema;
+  }
+  if (node.props.has("properties")) {
+    const raw = node.props.get("properties");
+    const resolved = raw && raw.t === "ident" ? resolveIdent(raw.name, ctx, depth) : raw;
+    if (resolved && resolved.t === "obj") {
+      schema.type = schema.type ?? "object";
+      schema.properties = objectSchema(resolved, ctx, depth).properties;
+    }
+  }
+  if (node.props.has("required")) {
+    const required = plainOf(node.props.get("required"));
+    if (Array.isArray(required)) schema.required = required;
+  } else if (schema.properties) {
+    schema.required = Object.keys(schema.properties);
+  }
+  if (schema.type === "object" && !schema.properties) schema.properties = {};
+  return schema;
+}
+
+function applyOptions(schema, optionsNode) {
+  for (const [key, valueNode] of optionsNode.props) {
+    if (key === "description") {
+      const value = plainOf(valueNode);
+      if (typeof value === "string") schema.description = value;
+      continue;
+    }
+    if (["format", "default", "minimum", "maximum", "minLength", "maxLength", "pattern"].includes(key)) {
+      const value = plainOf(valueNode);
+      if (value !== undefined) schema[key] = value;
+    }
+  }
+}
+
+function applyMethod(schema, chained) {
+  const [first] = chained.args ?? [];
+  switch (chained.name) {
+    case "describe": {
+      const value = plainOf(first);
+      if (typeof value === "string") schema.description = value;
+      return null;
+    }
+    case "optional":
+      return "optional";
+    case "default": {
+      const value = plainOf(first);
+      if (value !== undefined) schema.default = value;
+      return null;
+    }
+    case "min": {
+      const value = plainOf(first);
+      if (typeof value === "number") {
+        if (schema.type === "string") schema.minLength = value;
+        else schema.minimum = value;
+      }
+      return null;
+    }
+    case "max": {
+      const value = plainOf(first);
+      if (typeof value === "number") {
+        if (schema.type === "string") schema.maxLength = value;
+        else schema.maximum = value;
+      }
+      return null;
+    }
+    case "int":
+      schema.type = "integer";
+      return null;
+    case "uuid":
+      schema.format = "uuid";
+      return null;
+    default:
+      return null;
+  }
+}
+
+function plainOf(node, depth = 0) {
+  if (!node || depth > MAX_VALUE_DEPTH) return undefined;
+  switch (node.t) {
+    case "str":
+    case "num":
+    case "bool":
+      return node.value;
+    case "null":
+      return null;
+    case "arr": {
+      const values = node.items.map((item) => plainOf(item, depth + 1));
+      return values.some((value) => value === undefined) ? undefined : values;
+    }
+    case "obj": {
+      if (node.props.has("type") || node.props.has("properties")) return undefined;
+      const value = {};
+      for (const [key, item] of node.props) {
+        const plain = plainOf(item, depth + 1);
+        if (plain === undefined) return undefined;
+        value[key] = plain;
+      }
+      return value;
+    }
+    default:
+      return undefined;
+  }
+}
+
+// ---- bindings and identifier resolution ----------------------------------
+
+function collectBindings(src) {
+  const bindings = new Map();
+  const pattern = /(?:^|[;\n{}])\s*(?:export\s+)?(?:const|let|var)\s+([A-Za-z_$][\w$]*)\s*(?::[^=\n]{0,200})?=\s*/g;
+  for (const match of src.matchAll(pattern)) {
+    if (!bindings.has(match[1])) bindings.set(match[1], match.index + match[0].length);
+  }
+  return bindings;
+}
+
+function resolveIdent(name, ctx, depth = 0) {
+  if (depth > MAX_RESOLVE_DEPTH) return null;
+  if (ctx.parsed.has(name)) return ctx.parsed.get(name);
+  const start = ctx.bindings.get(name);
+  if (start === undefined) return null;
+  ctx.parsed.set(name, null);
+  const parsed = parseValue(ctx.src, start, 0);
+  const node = parsed ? parsed.node : null;
+  ctx.parsed.set(name, node);
+  return node;
+}
+
+// ---- value parser --------------------------------------------------------
+
+function parseValue(src, start, depth) {
+  if (depth > MAX_VALUE_DEPTH) return null;
+  const i = skipTrivia(src, start);
+  const ch = src[i];
+  if (ch === undefined) return null;
+  if (ch === "{") return parseObject(src, i, depth);
+  if (ch === "[") return parseArray(src, i, depth);
+  if (ch === '"' || ch === "'" || ch === "`") return parseString(src, i);
+  if (ch === "-" || (ch >= "0" && ch <= "9")) return parseNumber(src, i);
+  if (ch === "(") return tryArrow(src, i, depth);
+  const ident = readIdent(src, i);
+  if (ident) {
+    const arrow = tryArrow(src, i, depth);
+    if (arrow) return arrow;
+    if (!RESERVED.has(ident.name)) return parseChain(src, i, depth);
+  }
+  return null;
+}
+
+// `lazySchema(() => z.object({...}))` — the schema lives in the arrow body. A
+// block body is a function, not a value, so it becomes unknown rather than a
+// guess.
+function tryArrow(src, start, depth) {
+  let bodyFrom;
+  if (src[start] === "(") {
+    const close = matchParen(src, start);
+    if (close === null) return null;
+    bodyFrom = close + 1;
+  } else {
+    const ident = readIdent(src, start);
+    if (!ident) return null;
+    bodyFrom = ident.end;
+  }
+  const arrowAt = skipTrivia(src, bodyFrom);
+  if (!src.startsWith("=>", arrowAt)) return null;
+  const bodyStart = skipTrivia(src, arrowAt + 2);
+  if (src[bodyStart] === "{") {
+    const end = skipBraced(src, bodyStart + 1);
+    return end === null ? null : { node: { t: "unknown" }, end };
+  }
+  return parseValue(src, bodyStart, depth + 1);
+}
+
+function matchParen(src, start) {
+  let depth = 0;
+  let i = start;
+  while (i < src.length) {
+    const ch = src[i];
+    if (ch === '"' || ch === "'" || ch === "`") {
+      const parsed = parseString(src, i);
+      if (parsed) {
+        i = parsed.end;
+        continue;
+      }
+    }
+    if (ch === "(") depth += 1;
+    else if (ch === ")") {
+      depth -= 1;
+      if (depth === 0) return i;
+    }
+    i += 1;
+  }
+  return null;
+}
+
+function parseObject(src, start, depth) {
+  const props = new Map();
+  let i = skipTrivia(src, start + 1);
+  if (src[i] === "}") return { node: { t: "obj", props }, end: i + 1 };
+  for (;;) {
+    i = skipTrivia(src, i);
+    // A trailing comma leaves us on the closing brace.
+    if (src[i] === "}") return { node: { t: "obj", props }, end: i + 1 };
+    if (src.startsWith("...", i)) return null;
+    let key = null;
+    if (src[i] === '"' || src[i] === "'" || src[i] === "`") {
+      const parsed = parseString(src, i);
+      if (!parsed) return null;
+      key = parsed.node.value;
+      i = parsed.end;
+    } else {
+      const ident = readIdent(src, i);
+      if (ident) {
+        key = ident.name;
+        i = ident.end;
+      } else if (src[i] >= "0" && src[i] <= "9") {
+        const num = parseNumber(src, i);
+        if (!num) return null;
+        key = String(num.node.value);
+        i = num.end;
+      } else {
+
+      return null;
+      }
+    }
+    i = skipTrivia(src, i);
+    if (src[i] !== ":") {
+      // Property shorthand: `{ inputSchema, name, description }` refers to a
+      // local binding of the same name.
+      if (src[i] === "," || src[i] === "}") {
+        props.set(key, { t: "ident", name: key });
+        if (src[i] === ",") {
+          i += 1;
+          continue;
+        }
+        return { node: { t: "obj", props }, end: i + 1 };
+      }
+      // Accessors: `get inputSchema(): T { return inputSchema(); }`. Only a
+      // single-return body is taken; a conditional getter is left unknown
+      // rather than guessed from its first branch.
+      if ((key === "get" || key === "set") && /[A-Za-z_$]/.test(src[i] ?? "")) {
+        const accessor = readIdent(src, i);
+        const parenAt = skipTrivia(src, accessor.end);
+        if (src[parenAt] === "(") {
+          const params = parseArgs(src, parenAt);
+          if (!params) return null;
+          let braceAt = skipTrivia(src, params.end);
+          // `get inputSchema(): InputSchema {` — skip a simple return type.
+          if (src[braceAt] === ":") braceAt = findBrace(src, braceAt + 1);
+          if (src[braceAt] !== "{") return null;
+          const bodyEnd = skipBraced(src, braceAt + 1);
+          if (bodyEnd === null) return null;
+          props.set(accessor.name, getterValue(src, braceAt + 1, bodyEnd, depth));
+          i = skipTrivia(src, bodyEnd);
+          if (src[i] === ",") {
+            i += 1;
+            continue;
+          }
+          if (src[i] === "}") return { node: { t: "obj", props }, end: i + 1 };
+          return null;
+        }
+      }
+      // Shorthand methods and getters: `async call(args) { ... }`, `get x() {}`.
+      const shorthand = src[i] === "(" || src[i] === "<" || src[i] === "[" || /[A-Za-z_$]/.test(src[i] ?? "");
+      if (!shorthand) {
+      return null; }
+      const skipped = skipToValueEnd(src, i);
+      if (skipped === null) {
+      return null; }
+      props.set(key, { t: "unknown" });
+      i = skipped;
+      if (src[i] === ",") {
+        i += 1;
+        continue;
+      }
+      if (src[i] === "}") return { node: { t: "obj", props }, end: i + 1 };
+      return null;
+    }
+    const value = parseValue(src, i + 1, depth + 1);
+    const after = value ? skipTrivia(src, value.end) : -1;
+    if (value && (src[after] === "," || src[after] === "}")) {
+      props.set(key, value.node);
+      i = after;
+    } else {
+      // Unparseable or only partially parsed expression (`x => a || b`):
+      // keep the keys we understand and skip this value.
+      const skipped = skipToValueEnd(src, i + 1);
+      if (skipped === null) return null;
+      props.set(key, { t: "unknown" });
+      i = skipped;
+    }
+    if (src[i] === ",") {
+      i += 1;
+      continue;
+    }
+    if (src[i] === "}") return { node: { t: "obj", props }, end: i + 1 };
+    return null;
+  }
+}
+
+function findBrace(src, start) {
+  let i = start;
+  while (i < src.length) {
+    const ch = src[i];
+    if (ch === '"' || ch === "'" || ch === "`") {
+      const parsed = parseString(src, i);
+      if (parsed) {
+        i = parsed.end;
+        continue;
+      }
+    }
+    if (ch === "{") return i;
+    if (ch === ";" || ch === "}") return i;
+    i += 1;
+  }
+  return i;
+}
+
+function getterValue(src, start, end, depth) {
+  const body = src.slice(start, end);
+  const returns = body.match(/\breturn\b/g) ?? [];
+  if (returns.length !== 1) return { t: "unknown" };
+  const at = start + body.indexOf("return") + "return".length;
+  const parsed = parseValue(src, at, depth + 1);
+  return parsed ? parsed.node : { t: "unknown" };
+}
+
+function skipToValueEnd(src, start) {
+  let i = start;
+  let depth = 0;
+  while (i < src.length) {
+    const ch = src[i];
+    if (ch === '"' || ch === "'" || ch === "`") {
+      const parsed = parseString(src, i);
+      if (parsed) {
+        i = parsed.end;
+        continue;
+      }
+    }
+    if (ch === "/" && src[i + 1] === "/") {
+      const nl = src.indexOf("\n", i);
+      i = nl < 0 ? src.length : nl + 1;
+      continue;
+    }
+    if (ch === "(" || ch === "[" || ch === "{") depth += 1;
+    else if (ch === ")" || ch === "]") {
+      if (depth > 0) depth -= 1;
+    } else if (ch === "}") {
+      if (depth === 0) return i;
+      depth -= 1;
+    } else if (ch === "," && depth === 0) return i;
+    i += 1;
+  }
+  return null;
+}
+
+function parseArray(src, start, depth) {
+  const items = [];
+  let i = skipTrivia(src, start + 1);
+  if (src[i] === "]") return { node: { t: "arr", items }, end: i + 1 };
+  for (;;) {
+    const value = parseValue(src, i, depth + 1);
+    if (!value) return null;
+    items.push(value.node);
+    i = skipTrivia(src, value.end);
+    if (src[i] === ",") {
+      i = skipTrivia(src, i + 1);
+      if (src[i] === "]") return { node: { t: "arr", items }, end: i + 1 };
+      continue;
+    }
+    if (src[i] === "]") return { node: { t: "arr", items }, end: i + 1 };
+    return null;
+  }
+}
+
+function parseString(src, start) {
+  const quote = src[start];
+  if (quote === "`") return parseTemplate(src, start);
+  let i = start + 1;
+  let value = "";
+  while (i < src.length) {
+    const ch = src[i];
+    if (ch === "\\") {
+      const next = src[i + 1];
+      const map = { n: "\n", t: "\t", r: "\r", b: "\b", f: "\f", v: "\v", "0": "\0" };
+      value += map[next] ?? next ?? "";
+      i += 2;
+      continue;
+    }
+    if (ch === quote) return { node: { t: "str", value }, end: i + 1 };
+    if (ch === "\n") return null;
+    value += ch;
+    i += 1;
+  }
+  return null;
+}
+
+// Template literals appear inside handler bodies (`Slept for ${n} seconds`).
+// Their end must be found structurally: a naive scan confuses the closing and
+// opening backticks of two different templates and unbalances the object parse.
+function parseTemplate(src, start) {
+  let i = start + 1;
+  let value = "";
+  let dynamic = false;
+  while (i < src.length) {
+    const ch = src[i];
+    if (ch === "\\") {
+      value += src[i + 1] ?? "";
+      i += 2;
+      continue;
+    }
+    if (ch === "`") {
+      return dynamic ? { node: { t: "unknown" }, end: i + 1 } : { node: { t: "str", value }, end: i + 1 };
+    }
+    if (ch === "$" && src[i + 1] === "{") {
+      dynamic = true;
+      const end = skipBraced(src, i + 2);
+      if (end === null) return null;
+      i = end;
+      continue;
+    }
+    value += ch;
+    i += 1;
+  }
+  return null;
+}
+
+function skipBraced(src, start) {
+  let depth = 1;
+  let i = start;
+  while (i < src.length) {
+    const ch = src[i];
+    if (ch === '"' || ch === "'" || ch === "`") {
+      const parsed = parseString(src, i);
+      if (parsed) {
+        i = parsed.end;
+        continue;
+      }
+    }
+    if (ch === "{") depth += 1;
+    else if (ch === "}") {
+      depth -= 1;
+      if (depth === 0) return i + 1;
+    }
+    i += 1;
+  }
+  return null;
+}
+
+function parseNumber(src, start) {
+  // Numeric separators (`50_000`) are common in real tool defaults.
+  const match = /^-?\d[\d_]*(?:\.\d[\d_]*)?(?:e[+-]?\d+)?/i.exec(src.slice(start, start + 40));
+  if (!match) return null;
+  return { node: { t: "num", value: Number(match[0].replace(/_/g, "")) }, end: start + match[0].length };
+}
+
+function parseChain(src, start, depth) {
+  let i = start;
+  const segments = [];
+  const head = readIdent(src, i);
+  if (!head) return null;
+  segments.push(head.name);
+  i = head.end;
+  for (;;) {
+    const j = skipTrivia(src, i);
+    if (src[j] === ".") {
+      const next = readIdent(src, j + 1);
+      if (!next) break;
+      segments.push(next.name);
+      i = next.end;
+      continue;
+    }
+    if (src[j] === "(") {
+      const parsed = parseArgs(src, j);
+      if (!parsed) return null;
+      const calls = [];
+      let end = parsed.end;
+      for (;;) {
+        const k = skipTrivia(src, end);
+        if (src[k] !== ".") break;
+        const method = readIdent(src, k + 1);
+        if (!method) break;
+        const l = skipTrivia(src, method.end);
+        if (src[l] !== "(") break;
+        const methodArgs = parseArgs(src, l);
+        if (!methodArgs) break;
+        calls.push({ name: method.name, args: methodArgs.items });
+        end = methodArgs.end;
+      }
+      return { node: { t: "call", callee: segments.join("."), args: parsed.items, calls }, end };
+    }
+    break;
+  }
+  return { node: { t: "ident", name: segments.join(".") }, end: i };
+}
+
+function parseArgs(src, start) {
+  const items = [];
+  let i = skipTrivia(src, start + 1);
+  if (src[i] === ")") return { items, end: i + 1 };
+  for (;;) {
+    let value = parseValue(src, i, 0);
+    let after = value ? skipTrivia(src, value.end) : -1;
+    if (value) {
+      // TypeScript wrappers: `{...} satisfies ToolDef<...>` and `x as T`.
+      if (isTypeWrapperAt(src, after)) {
+        const skipped = skipTypeExpression(src, after);
+        if (skipped === null) return null;
+        after = skipped;
+      }
+      if (src[after] === "," || src[after] === ")") {
+        items.push(value.node);
+        i = after;
+      } else {
+        // Binary expressions (`a || b`, `a ?? b`) parse only their first
+        // operand; treat the whole argument as unknown instead.
+        value = null;
+      }
+    }
+    if (!value) {
+      const skipped = skipArgument(src, i);
+      if (skipped === null) return null;
+      items.push(null);
+      i = skipped;
+    }
+    if (src[i] === ",") {
+      i = skipTrivia(src, i + 1);
+      continue;
+    }
+    if (src[i] === ")") return { items, end: i + 1 };
+    return null;
+  }
+}
+
+function isTypeWrapperAt(src, i) {
+  return (
+    (src.startsWith("satisfies", i) && !/[\w$]/.test(src[i + 9] ?? "")) ||
+    (src.startsWith("as", i) && /[\s<]/.test(src[i + 2] ?? ""))
+  );
+}
+
+// Skips an annotation after satisfies/as. Generic argument commas are not
+// argument separators, so `<...>` is tracked as depth.
+function skipTypeExpression(src, start) {
+  let i = start;
+  let paren = 0;
+  let angle = 0;
+  while (i < src.length) {
+    const ch = src[i];
+    if (ch === '"' || ch === "'" || ch === "`") {
+      const parsed = parseString(src, i);
+      if (parsed) {
+        i = parsed.end;
+        continue;
+      }
+    }
+    if (ch === "(" || ch === "[" || ch === "{") paren += 1;
+    else if (ch === ")" || ch === "]" || ch === "}") {
+      if (paren === 0) return i;
+      paren -= 1;
+    } else if (ch === "<") angle += 1;
+    else if (ch === ">" && angle > 0) angle -= 1;
+    else if (ch === "," && paren === 0 && angle === 0) return i;
+    i += 1;
+  }
+  return null;
+}
+
+// A handler argument (`async (args) => {...}`) is not a value we can parse; step
+// over it so one unparseable parameter does not discard the whole definition.
+function skipArgument(src, start) {
+  let i = start;
+  let depth = 0;
+  while (i < src.length) {
+    const ch = src[i];
+    if (ch === '"' || ch === "'" || ch === "`") {
+      const parsed = parseString(src, i);
+      if (parsed) {
+        i = parsed.end;
+        continue;
+      }
+    }
+    if (ch === "(" || ch === "[" || ch === "{") depth += 1;
+    else if (ch === ")" || ch === "]" || ch === "}") {
+      if (depth === 0) return ch === ")" ? i : null;
+      depth -= 1;
+    } else if (ch === "," && depth === 0) return i;
+    i += 1;
+  }
+  return null;
+}
+
+function readIdent(src, start) {
+  const match = /^[A-Za-z_$][\w$]*/.exec(src.slice(start, start + 120));
+  if (!match) return null;
+  return { name: match[0], end: start + match[0].length };
+}
+
+function skipTrivia(src, start) {
+  let i = start;
+  while (i < src.length) {
+    const ch = src[i];
+    if (ch === " " || ch === "\t" || ch === "\r" || ch === "\n") {
+      i += 1;
+      continue;
+    }
+    if (ch === "/" && src[i + 1] === "/") {
+      const nl = src.indexOf("\n", i);
+      i = nl < 0 ? src.length : nl + 1;
+      continue;
+    }
+    if (ch === "/" && src[i + 1] === "*") {
+      const close = src.indexOf("*/", i + 2);
+      i = close < 0 ? src.length : close + 2;
+      continue;
+    }
+    return i;
+  }
+  return i;
+}

package/lib/mcp-server.js

@@ -57,6 +57,7 @@ export async function startMcpServer() {
             source: mcpServer.source,
             origin: "mcp-probe",
             confidence: "high",
+            replaceExisting: true,
             note: `server ${mcpServer.name}`
           });
           report.probes.push({ server: mcpServer.name, ok: true, toolCount: added.length, tools: added });

package/lib/scan.js

@@ -1,5 +1,6 @@
 import { readdir, readFile, stat } from "node:fs/promises";
 import { basename, extname, join, relative, resolve } from "node:path";
+import { extractJsTools } from "./js-tools.js";
 
 const SKIP_DIRS = new Set([
   "node_modules",
@@ -19,7 +20,10 @@ const SKIP_DIRS = new Set([
 ]);
 
 const MAX_DEPTH = 6;
-const MAX_FILES = 5000;
+// 5k was too small: a real agent monorepo (openclaw) has 16k files after the
+// skip list, so the walk stopped a third of the way in and reported 26 of its
+// 53 skills. Silent partial results are worse than a slow scan.
+const MAX_FILES = 25000;
 const MAX_FILE_BYTES = 200_000;
 
 const FUNCTION_SPEC_NAMES = new Set([
@@ -81,7 +85,8 @@ const JS_FRAMEWORKS = {
 export async function scanProject(root = ".") {
   const absRoot = resolve(root);
   const files = [];
-  await walk(absRoot, 0, files);
+  const state = { truncated: false };
+  await walk(absRoot, 0, files, state);
 
   const report = {
     root: absRoot,
@@ -97,7 +102,16 @@ export async function scanProject(root = ".") {
     warnings: []
   };
 
+  if (state.truncated) {
+    report.warnings.push(`扫描在 ${MAX_FILES} 个文件处截断，结果可能不完整；可把 --root 指向子目录`);
+  }
+
   const seenTools = new Set();
+  // Tool names are frequently constants defined in a sibling module
+  // (`name: CONFIG_TOOL_NAME`). Files are visited in path order, so the
+  // constant may not be known yet; such files are retried after the walk.
+  const constIndex = new Map();
+  const retryFiles = new Set();
   const dependencies = new Set();
   const frameworks = new Set();
 
@@ -224,6 +238,23 @@ export async function scanProject(root = ".") {
           });
         }
       }
+      if (!name.endsWith(".py") && !TEST_PATH.test(rel)) {
+        for (const match of text.matchAll(/(?:export\s+)?const\s+([A-Z][A-Z0-9_]{2,})\s*=\s*(['"])([^'"\n]{1,200})\2/g)) {
+          if (!constIndex.has(match[1])) constIndex.set(match[1], match[3]);
+        }
+        const jsTools = extractJsTools(text, {
+          externals: constIndex,
+          onUnresolvedName: () => retryFiles.add(file)
+        });
+        if (jsTools.length > 0) {
+          addTools(report, jsTools, {
+            source: rel,
+            origin: "code-literal",
+            confidence: "medium",
+            note: "从 JS/TS 声明或 TypeBox/zod schema 静态提取（不执行代码），需人工确认"
+          });
+        }
+      }
       if (TEST_PATH.test(rel)) continue;
       // Only matches that carry a tool name are actionable; a bare `input_schema:`
       // in API plumbing is not a tool and only inflates the report.
@@ -249,19 +280,46 @@ export async function scanProject(root = ".") {
     }
   }
 
+  if (retryFiles.size > 0 && constIndex.size > 0) {
+    for (const file of retryFiles) {
+      const text = await readText(file, report);
+      if (!text) continue;
+      const tools = extractJsTools(text, { externals: constIndex });
+      if (tools.length === 0) continue;
+      addTools(report, tools, {
+        source: relative(absRoot, file),
+        origin: "code-literal",
+        confidence: "medium",
+        note: "名称常量在其他模块定义，扫描完整仓库后再解析"
+      });
+    }
+  }
+
   report.dependencies = [...dependencies].sort();
   report.frameworks = [...frameworks].sort();
   return report;
 }
 
-export function addTools(report, tools, { source, origin, confidence = "high", note, meta }) {
-  const existing = new Set(report.tools.map((tool) => tool.name));
+export function addTools(
+  report,
+  tools,
+  { source, origin, confidence = "high", note, meta, replaceExisting = false }
+) {
+  const index = new Map(report.tools.map((tool, position) => [tool.name, position]));
   const added = [];
   for (const tool of tools) {
-    if (!tool?.name || existing.has(tool.name)) continue;
-    existing.add(tool.name);
-    report.tools.push({ ...tool, source, origin, ...(meta ?? {}) });
-    added.push(tool.name);
+    if (!tool?.name) continue;
+    const existing = index.get(tool.name);
+    if (existing === undefined) {
+      index.set(tool.name, report.tools.length);
+      report.tools.push({ ...tool, source, origin, ...(meta ?? {}) });
+      added.push(tool.name);
+    } else if (replaceExisting) {
+      // A probed tool is the runtime truth; a static guess with the same name
+      // should be replaced, not silently kept.
+      report.tools[existing] = { ...tool, source, origin, ...(meta ?? {}) };
+      added.push(tool.name);
+    }
   }
   report.sources.push({
     type: origin,
@@ -387,8 +445,11 @@ export function formatScanSummary(report) {
   return lines.join("\n");
 }
 
-async function walk(dir, depth, files) {
-  if (depth > MAX_DEPTH || files.length >= MAX_FILES) return;
+async function walk(dir, depth, files, state) {
+  if (depth > MAX_DEPTH || files.length >= MAX_FILES) {
+    if (files.length >= MAX_FILES) state.truncated = true;
+    return;
+  }
   let entries;
   try {
     entries = await readdir(dir, { withFileTypes: true });
@@ -397,11 +458,14 @@ async function walk(dir, depth, files) {
   }
   entries.sort((a, b) => a.name.localeCompare(b.name));
   for (const entry of entries) {
-    if (files.length >= MAX_FILES) return;
+    if (files.length >= MAX_FILES) {
+      state.truncated = true;
+      return;
+    }
     const full = join(dir, entry.name);
     if (entry.isDirectory()) {
       if (SKIP_DIRS.has(entry.name)) continue;
-      await walk(full, depth + 1, files);
+      await walk(full, depth + 1, files, state);
     } else if (entry.isFile()) {
       files.push(full);
     }

package/lib/suite.js

@@ -3,10 +3,10 @@ import { dirname, resolve } from "node:path";
 
 const EMPTY_SCHEMA = { type: "object", properties: {} };
 
-export async function loadSuite(suitePath, { allowEmptyCases = false } = {}) {
+export async function loadSuite(suitePath, { allowEmptyCases = false, allowEmptyTools = false } = {}) {
   const abs = resolve(suitePath);
   const raw = JSON.parse(await readFile(abs, "utf8"));
-  if (!Array.isArray(raw.tools) || raw.tools.length === 0) {
+  if (!Array.isArray(raw.tools) || (raw.tools.length === 0 && !allowEmptyTools)) {
     throw new Error(`suite ${abs}: "tools" must be a non-empty array`);
   }
   if (!Array.isArray(raw.cases) || (raw.cases.length === 0 && !allowEmptyCases)) {

package/mf.js

@@ -344,6 +344,7 @@ export async function runScan(options, io = console) {
         source: server.source,
         origin: "mcp-probe",
         confidence: "high",
+        replaceExisting: true,
         note: `server ${server.name}`,
         meta: { provider: server.name }
       });
@@ -408,8 +409,11 @@ export async function runCaseAdd(options, io = console) {
 
 export async function runCoverage(options, io = console) {
   // A scan draft has no cases yet; "what is still missing" is exactly the view
-  // a user needs right after scanning.
-  const suite = await loadSuite(resolve(options.suite ?? DEFAULT_SUITE), { allowEmptyCases: true });
+  // a user needs right after scanning. A skills-only suite is also valid.
+  const suite = await loadSuite(resolve(options.suite ?? DEFAULT_SUITE), {
+    allowEmptyCases: true,
+    allowEmptyTools: true
+  });
   const report = computeCoverage(suite);
   io.log(options.json ? JSON.stringify(report, null, 2) : formatCoverage(report));
   return 0;
@@ -422,7 +426,10 @@ export async function runSandboxCheck(options, io = console) {
 }
 
 export async function runLint(options, io = console) {
-  const suite = await loadSuite(resolve(options.suite ?? DEFAULT_SUITE), { allowEmptyCases: true });
+  const suite = await loadSuite(resolve(options.suite ?? DEFAULT_SUITE), {
+    allowEmptyCases: true,
+    allowEmptyTools: true
+  });
   const report = lintSuite(suite);
   io.log(options.json ? JSON.stringify(report, null, 2) : formatLint(report));
   return report.counts.error > 0 ? 1 : 0;

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "modelfunction-cli",
-  "version": "0.3.0",
+  "version": "0.4.0",
   "type": "module",
   "license": "MIT",
   "description": "mf — function-calling accuracy scorecard for AI agents",