gitnexus 1.6.3-rc.40 → 1.6.3-rc.42

package/README.md

@@ -155,6 +155,7 @@ gitnexus analyze --force         # Force full re-index
 gitnexus analyze --embeddings    # Enable embedding generation (slower, better search)
 gitnexus analyze --skip-agents-md  # Preserve custom AGENTS.md/CLAUDE.md gitnexus section edits
 gitnexus analyze --verbose       # Log skipped files when parsers are unavailable
+gitnexus analyze --max-file-size 1024  # Skip files larger than N KB (default: 512, cap: 32768)
 gitnexus mcp                     # Start MCP server (stdio) — serves all indexed repos
 gitnexus serve                   # Start local HTTP server (multi-repo) for web UI
 gitnexus index                   # Register an existing .gitnexus/ folder into the global registry
@@ -307,6 +308,21 @@ echo "vendor/" >> .gitnexusignore
 echo "dist/" >> .gitnexusignore
 ```
 
+### Large files are being skipped
+
+By default the walker skips files larger than **512 KB** (see log line `Skipped N large files (>512KB)`). Raise the threshold via either the CLI flag or the environment variable — both accept a value in **KB**:
+
+```bash
+# CLI flag (takes precedence over the env var)
+npx gitnexus analyze --max-file-size 2048     # skip only files > 2 MB
+
+# Environment variable (persists across commands)
+export GITNEXUS_MAX_FILE_SIZE=2048
+npx gitnexus analyze
+```
+
+Values above **32768 KB (32 MB)** are clamped to the tree-sitter parser ceiling; invalid values fall back to the 512 KB default with a one-time warning. When an override is active, `analyze` prints the effective threshold in its startup banner (e.g. `GITNEXUS_MAX_FILE_SIZE: effective threshold 2048KB (default 512KB)`).
+
 ## Privacy
 
 - All processing happens locally on your machine

package/dist/_shared/scope-resolution/parsed-file.d.ts

@@ -37,6 +37,18 @@
  * `localDefs`. A `ParsedFile` is trivially convertible to a `FinalizeFile`
  * by picking those four fields, so the finalize orchestrator threads
  * ParsedFile through to the shared algorithm without shape-shifting.
+ *
+ * ## Source-of-truth invariant
+ *
+ * `ParsedFile` is the single semantic model consumed by both the legacy
+ * DAG (`gitnexus/src/core/ingestion/` outside `scope-resolution/`) and
+ * the scope-resolution pipeline (`gitnexus/src/core/ingestion/scope-resolution/`).
+ * Downstream passes MUST NOT build a parallel parse representation; if
+ * a pass needs AST-level facts that `ParsedFile` doesn't expose, it
+ * should reuse the orchestrator's `treeCache` rather than re-invoke
+ * `parser.parse(...)` on its own. See the
+ * `ScopeResolver` contract (`gitnexus/src/core/ingestion/scope-resolution/contract/scope-resolver.ts`)
+ * for the full list of invariants downstream consumers rely on.
  */
 import type { Scope, ScopeId } from './types.js';
 import type { ParsedImport } from './types.js';

package/dist/_shared/scope-resolution/parsed-file.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"parsed-file.d.ts","sourceRoot":"","sources":["../../src/scope-resolution/parsed-file.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;GAuCG;AAEH,OAAO,KAAK,EAAE,KAAK,EAAE,OAAO,EAAE,MAAM,YAAY,CAAC;AACjD,OAAO,KAAK,EAAE,YAAY,EAAE,MAAM,YAAY,CAAC;AAC/C,OAAO,KAAK,EAAE,gBAAgB,EAAE,MAAM,wBAAwB,CAAC;AAC/D,OAAO,KAAK,EAAE,aAAa,EAAE,MAAM,qBAAqB,CAAC;AAEzD,MAAM,WAAW,UAAU;IACzB,QAAQ,CAAC,QAAQ,EAAE,MAAM,CAAC;IAC1B,oDAAoD;IACpD,QAAQ,CAAC,WAAW,EAAE,OAAO,CAAC;IAC9B;;;;OAIG;IACH,QAAQ,CAAC,MAAM,EAAE,SAAS,KAAK,EAAE,CAAC;IAClC,QAAQ,CAAC,aAAa,EAAE,SAAS,YAAY,EAAE,CAAC;IAChD;;;;OAIG;IACH,QAAQ,CAAC,SAAS,EAAE,SAAS,gBAAgB,EAAE,CAAC;IAChD,QAAQ,CAAC,cAAc,EAAE,SAAS,aAAa,EAAE,CAAC;CACnD"}
+{"version":3,"file":"parsed-file.d.ts","sourceRoot":"","sources":["../../src/scope-resolution/parsed-file.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;GAmDG;AAEH,OAAO,KAAK,EAAE,KAAK,EAAE,OAAO,EAAE,MAAM,YAAY,CAAC;AACjD,OAAO,KAAK,EAAE,YAAY,EAAE,MAAM,YAAY,CAAC;AAC/C,OAAO,KAAK,EAAE,gBAAgB,EAAE,MAAM,wBAAwB,CAAC;AAC/D,OAAO,KAAK,EAAE,aAAa,EAAE,MAAM,qBAAqB,CAAC;AAEzD,MAAM,WAAW,UAAU;IACzB,QAAQ,CAAC,QAAQ,EAAE,MAAM,CAAC;IAC1B,oDAAoD;IACpD,QAAQ,CAAC,WAAW,EAAE,OAAO,CAAC;IAC9B;;;;OAIG;IACH,QAAQ,CAAC,MAAM,EAAE,SAAS,KAAK,EAAE,CAAC;IAClC,QAAQ,CAAC,aAAa,EAAE,SAAS,YAAY,EAAE,CAAC;IAChD;;;;OAIG;IACH,QAAQ,CAAC,SAAS,EAAE,SAAS,gBAAgB,EAAE,CAAC;IAChD,QAAQ,CAAC,cAAc,EAAE,SAAS,aAAa,EAAE,CAAC;CACnD"}

package/dist/_shared/scope-resolution/parsed-file.js

@@ -37,6 +37,18 @@
  * `localDefs`. A `ParsedFile` is trivially convertible to a `FinalizeFile`
  * by picking those four fields, so the finalize orchestrator threads
  * ParsedFile through to the shared algorithm without shape-shifting.
+ *
+ * ## Source-of-truth invariant
+ *
+ * `ParsedFile` is the single semantic model consumed by both the legacy
+ * DAG (`gitnexus/src/core/ingestion/` outside `scope-resolution/`) and
+ * the scope-resolution pipeline (`gitnexus/src/core/ingestion/scope-resolution/`).
+ * Downstream passes MUST NOT build a parallel parse representation; if
+ * a pass needs AST-level facts that `ParsedFile` doesn't expose, it
+ * should reuse the orchestrator's `treeCache` rather than re-invoke
+ * `parser.parse(...)` on its own. See the
+ * `ScopeResolver` contract (`gitnexus/src/core/ingestion/scope-resolution/contract/scope-resolver.ts`)
+ * for the full list of invariants downstream consumers rely on.
  */
 export {};
 //# sourceMappingURL=parsed-file.js.map

package/dist/_shared/scope-resolution/parsed-file.js.map

@@ -1 +1 @@
-{"version":3,"file":"parsed-file.js","sourceRoot":"","sources":["../../src/scope-resolution/parsed-file.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;GAuCG"}
+{"version":3,"file":"parsed-file.js","sourceRoot":"","sources":["../../src/scope-resolution/parsed-file.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;GAmDG"}

package/dist/_shared/scope-resolution/reference-site.d.ts

@@ -63,5 +63,13 @@ export interface ReferenceSite {
     };
     /** Argument count at the call site; used by `provider.arityCompatibility`. */
     readonly arity?: number;
+    /**
+     * Inferred argument types at the call site, one per argument. An
+     * empty-string entry means "unknown" — consumers narrowing overload
+     * candidates treat unknown as any-match. Populated by languages
+     * that can derive types from literals / constructor expressions
+     * (C#: `42` → `'int'`, `"alice"` → `'string'`).
+     */
+    readonly argumentTypes?: readonly string[];
 }
 //# sourceMappingURL=reference-site.d.ts.map

package/dist/_shared/scope-resolution/reference-site.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"reference-site.d.ts","sourceRoot":"","sources":["../../src/scope-resolution/reference-site.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;;GAqBG;AAEH,OAAO,KAAK,EAAE,KAAK,EAAE,OAAO,EAAE,MAAM,YAAY,CAAC;AAEjD;;;;;;GAMG;AACH,MAAM,MAAM,aAAa,GACrB,MAAM,GACN,MAAM,GACN,OAAO,GACP,gBAAgB,GAChB,UAAU,GACV,YAAY,CAAC;AAEjB;;;;;;;;;GASG;AACH,MAAM,MAAM,QAAQ,GAAG,MAAM,GAAG,QAAQ,GAAG,aAAa,GAAG,OAAO,CAAC;AAEnE,MAAM,WAAW,aAAa;IAC5B,uEAAuE;IACvE,QAAQ,CAAC,IAAI,EAAE,MAAM,CAAC;IACtB,2CAA2C;IAC3C,QAAQ,CAAC,OAAO,EAAE,KAAK,CAAC;IACxB;;;;OAIG;IACH,QAAQ,CAAC,OAAO,EAAE,OAAO,CAAC;IAC1B,QAAQ,CAAC,IAAI,EAAE,aAAa,CAAC;IAC7B,kCAAkC;IAClC,QAAQ,CAAC,QAAQ,CAAC,EAAE,QAAQ,CAAC;IAC7B;;;OAGG;IACH,QAAQ,CAAC,gBAAgB,CAAC,EAAE;QAAE,QAAQ,CAAC,IAAI,EAAE,MAAM,CAAA;KAAE,CAAC;IACtD,8EAA8E;IAC9E,QAAQ,CAAC,KAAK,CAAC,EAAE,MAAM,CAAC;CACzB"}
+{"version":3,"file":"reference-site.d.ts","sourceRoot":"","sources":["../../src/scope-resolution/reference-site.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;;GAqBG;AAEH,OAAO,KAAK,EAAE,KAAK,EAAE,OAAO,EAAE,MAAM,YAAY,CAAC;AAEjD;;;;;;GAMG;AACH,MAAM,MAAM,aAAa,GACrB,MAAM,GACN,MAAM,GACN,OAAO,GACP,gBAAgB,GAChB,UAAU,GACV,YAAY,CAAC;AAEjB;;;;;;;;;GASG;AACH,MAAM,MAAM,QAAQ,GAAG,MAAM,GAAG,QAAQ,GAAG,aAAa,GAAG,OAAO,CAAC;AAEnE,MAAM,WAAW,aAAa;IAC5B,uEAAuE;IACvE,QAAQ,CAAC,IAAI,EAAE,MAAM,CAAC;IACtB,2CAA2C;IAC3C,QAAQ,CAAC,OAAO,EAAE,KAAK,CAAC;IACxB;;;;OAIG;IACH,QAAQ,CAAC,OAAO,EAAE,OAAO,CAAC;IAC1B,QAAQ,CAAC,IAAI,EAAE,aAAa,CAAC;IAC7B,kCAAkC;IAClC,QAAQ,CAAC,QAAQ,CAAC,EAAE,QAAQ,CAAC;IAC7B;;;OAGG;IACH,QAAQ,CAAC,gBAAgB,CAAC,EAAE;QAAE,QAAQ,CAAC,IAAI,EAAE,MAAM,CAAA;KAAE,CAAC;IACtD,8EAA8E;IAC9E,QAAQ,CAAC,KAAK,CAAC,EAAE,MAAM,CAAC;IACxB;;;;;;OAMG;IACH,QAAQ,CAAC,aAAa,CAAC,EAAE,SAAS,MAAM,EAAE,CAAC;CAC5C"}

package/dist/cli/index.js

@@ -30,7 +30,9 @@ program
     .option('--max-file-size <kb>', 'Skip files larger than this (KB). Default: 512. Hard cap: 32768 (tree-sitter limit).')
     .addHelpText('after', '\nEnvironment variables:\n' +
     '  GITNEXUS_NO_GITIGNORE=1   Skip .gitignore parsing (still reads .gitnexusignore)\n' +
-    '  GITNEXUS_MAX_FILE_SIZE=N  Override large-file skip threshold (KB). Default 512, max 32768.')
+    '  GITNEXUS_MAX_FILE_SIZE=N  Override large-file skip threshold (KB). Default 512, max 32768.\n' +
+    '\nTip: `.gitnexusignore` supports `.gitignore`-style negation. Add e.g.\n' +
+    '     `!__tests__/` to index a directory that is auto-filtered by default (#771).')
     .action(createLazyAction(() => import('./analyze.js'), 'analyzeCommand'));
 program
     .command('index [path...]')

package/dist/config/ignore-service.d.ts

@@ -19,6 +19,15 @@ export declare const loadIgnoreRules: (repoPath: string, options?: IgnoreOptions
  *
  * Returns an IgnoreLike object for glob's `ignore` option,
  * enabling directory-level pruning during traversal.
+ *
+ * Precedence (#771): user's `.gitnexusignore` negation patterns take
+ * priority over the hardcoded list, matching `.gitignore` semantics.
+ * An explicit `!pattern` rule unignores descendants even when they
+ * would otherwise be blocked by DEFAULT_IGNORE_LIST — UNLESS a more
+ * specific rule in the same file re-ignores a subset (e.g.
+ * `!__tests__/` paired with `__tests__/generated/` blocks the child
+ * while leaving the parent negated). Last-match-wins is enforced by
+ * consulting `ig.ignores(rel)` after `hasExplicitUnignore`.
  */
 export declare const createIgnoreFilter: (repoPath: string, options?: IgnoreOptions) => Promise<{
     ignored(p: Path): boolean;

package/dist/config/ignore-service.js

@@ -245,16 +245,20 @@ const IGNORED_FILES = new Set([
     '.env.test',
     '.env.example',
 ]);
-// NOTE: Negation patterns in .gitnexusignore (e.g. `!vendor/`) cannot override
-// entries in DEFAULT_IGNORE_LIST — this is intentional. The hardcoded list protects
-// against indexing directories that are almost never source code (node_modules, .git, etc.).
-// Users who need to include such directories should remove them from the hardcoded list.
+// The hardcoded DEFAULT_IGNORE_LIST is the "safety net" default: directories
+// that are almost never source code (node_modules, .git, dist, __tests__,
+// etc.). Users who legitimately need to index one of these can negate the
+// hardcoded rule via a `!pattern` line in `.gitnexusignore` (#771) — same
+// semantics as `.gitignore` negation. That override is applied in
+// `createIgnoreFilter` below; `shouldIgnorePath` itself stays a pure
+// hardcoded-list check so its callers (wiki generator, tests) get
+// deterministic results independent of per-repo config.
 export const shouldIgnorePath = (filePath) => {
     const normalizedPath = filePath.replace(/\\/g, '/');
     const parts = normalizedPath.split('/');
     const fileName = parts[parts.length - 1];
     const fileNameLower = fileName.toLowerCase();
-    // Check if any path segment is in ignore list
+    // Check if any path segment is in the hardcoded ignore list.
     for (const part of parts) {
         if (DEFAULT_IGNORE_LIST.has(part)) {
             return true;
@@ -320,6 +324,44 @@ export const loadIgnoreRules = async (repoPath, options) => {
     }
     return hasRules ? ig : null;
 };
+/**
+ * Walk ancestor segments of `rel` and check whether `.gitnexusignore`
+ * (or `.gitignore`) contains an explicit `!pattern` negation that
+ * applies. Returns true as soon as any segment — or the path itself —
+ * is matched by a negation rule.
+ *
+ * Why this exists (#771): the hardcoded DEFAULT_IGNORE_LIST would
+ * otherwise block indexing of directories like `__tests__/` even when
+ * the user has an explicit `!__tests__/` line in `.gitnexusignore`.
+ * Mirroring `.gitignore` negation semantics: a user's explicit
+ * unignore of a parent directory implicitly unignores everything
+ * underneath, so we walk the ancestor chain rather than only testing
+ * the leaf.
+ *
+ * The `ignore` package's `test(path)` returns `{ignored, unignored}`;
+ * `unignored: true` is the "a negation rule matched this path"
+ * signal. Children of a negated directory return
+ * `{ignored: false, unignored: false}` on a direct test, which is why
+ * we also walk the ancestors here.
+ */
+const hasExplicitUnignore = (ig, rel) => {
+    // Direct match on the path (as a file).
+    if (ig.test(rel).unignored)
+        return true;
+    // Direct match on the path treated as a directory — `!dir/` matches
+    // here when rel is the directory itself.
+    if (ig.test(rel + '/').unignored)
+        return true;
+    // Walk ancestor segments. `!parent/` should propagate to every
+    // descendant the same way `.gitignore` negation propagates.
+    const parts = rel.split('/');
+    for (let i = parts.length - 1; i > 0; i--) {
+        const ancestor = parts.slice(0, i).join('/') + '/';
+        if (ig.test(ancestor).unignored)
+            return true;
+    }
+    return false;
+};
 /**
  * Create a glob-compatible ignore filter combining:
  * - .gitignore / .gitnexusignore patterns (via `ignore` package)
@@ -327,6 +369,15 @@ export const loadIgnoreRules = async (repoPath, options) => {
  *
  * Returns an IgnoreLike object for glob's `ignore` option,
  * enabling directory-level pruning during traversal.
+ *
+ * Precedence (#771): user's `.gitnexusignore` negation patterns take
+ * priority over the hardcoded list, matching `.gitignore` semantics.
+ * An explicit `!pattern` rule unignores descendants even when they
+ * would otherwise be blocked by DEFAULT_IGNORE_LIST — UNLESS a more
+ * specific rule in the same file re-ignores a subset (e.g.
+ * `!__tests__/` paired with `__tests__/generated/` blocks the child
+ * while leaving the parent negated). Last-match-wins is enforced by
+ * consulting `ig.ignores(rel)` after `hasExplicitUnignore`.
  */
 export const createIgnoreFilter = async (repoPath, options) => {
     const ig = await loadIgnoreRules(repoPath, options);
@@ -337,6 +388,15 @@ export const createIgnoreFilter = async (repoPath, options) => {
             const rel = p.relative();
             if (!rel)
                 return false;
+            // User's .gitnexusignore negation takes precedence over hardcoded
+            // rules (#771). If any ancestor or the path itself was explicitly
+            // unignored AND no more-specific rule re-ignores this exact path,
+            // allow it through. The `!ig.ignores(rel)` guard matches
+            // .gitignore's last-match-wins semantics: `!__tests__/` followed
+            // by `__tests__/generated/` negates the parent but still blocks
+            // the re-ignored child.
+            if (ig && hasExplicitUnignore(ig, rel) && !ig.ignores(rel))
+                return false;
             // Check .gitignore / .gitnexusignore patterns
             if (ig && ig.ignores(rel))
                 return true;
@@ -344,10 +404,20 @@ export const createIgnoreFilter = async (repoPath, options) => {
             return shouldIgnorePath(rel);
         },
         childrenIgnored(p) {
-            // Fast path: check directory name against hardcoded list.
             // Note: dot-directories (.git, .vscode, etc.) are primarily excluded by
-            // glob's `dot: false` option in filesystem-walker.ts. This check is
-            // defense-in-depth — do not remove `dot: false` assuming this covers it.
+            // glob's `dot: false` option in filesystem-walker.ts. The hardcoded
+            // list check below is defense-in-depth — do not remove `dot: false`
+            // assuming this covers it.
+            const rel = p.relative();
+            // User's .gitnexusignore negation takes precedence (#771) — if the
+            // user explicitly unignored this directory or any ancestor via a
+            // !pattern rule, allow descent even if the directory name is in
+            // DEFAULT_IGNORE_LIST. The `!ig.ignores(rel + '/')` guard keeps
+            // last-match-wins: `!__tests__/` + `__tests__/generated/` still
+            // blocks descent into `__tests__/generated/`.
+            if (ig && rel && hasExplicitUnignore(ig, rel) && !ig.ignores(rel + '/'))
+                return false;
+            // Hardcoded list: block descent into well-known noise directories.
             if (DEFAULT_IGNORE_LIST.has(p.name))
                 return true;
             // Check against .gitignore / .gitnexusignore patterns.
@@ -358,11 +428,8 @@ export const createIgnoreFilter = async (repoPath, options) => {
             // Bare-name patterns (e.g. `local`) still match `local/` per gitignore spec:
             // the `ignore` package normalizes `dir` and `dir/` to match directories.
             // See: https://github.com/kaelzhang/node-ignore#2-filenames-and-dirnames
-            if (ig) {
-                const rel = p.relative();
-                if (rel && ig.ignores(rel + '/'))
-                    return true;
-            }
+            if (ig && rel && ig.ignores(rel + '/'))
+                return true;
             return false;
         },
     };

package/dist/core/ingestion/language-provider.d.ts

@@ -224,8 +224,9 @@ interface LanguageProviderConfig {
     readonly builtInNames?: ReadonlySet<string>;
     /**
      * Emit scope captures from raw source, **pre-grouped per tree-sitter
-     * query match**. Tree-sitter-based providers run a `scopes.scm` query
-     * and emit one `CaptureMatch` per query match; standalone providers
+     * query match**. Tree-sitter-based providers run a scope query
+     * (embedded as a string constant in each language's `query.ts`) and
+     * emit one `CaptureMatch` per query match; standalone providers
      * (COBOL) emit matches from a regex tagger. The return shape is
      * parser-agnostic: the central `ScopeExtractor` consumes
      * `CaptureMatch[]` without knowing which parser produced them.

package/dist/core/ingestion/languages/csharp/accessor-unwrap.d.ts

@@ -0,0 +1,21 @@
+/**
+ * C# collection-accessor unwrapping.
+ *
+ * When the compound-receiver resolver encounters a trailing
+ * `.Values` / `.Keys` on a dotted member-access chain, it calls the
+ * provider's `unwrapCollectionAccessor` hook to find the element
+ * type. This module supplies the C# implementation — recognizing
+ * Dictionary-family generics and returning the value or key type.
+ *
+ * Other languages (Python, Java, TypeScript) use method-call syntax
+ * for the same access (`.values()` / `.keys()`), which the compound-
+ * receiver's call-expression branch already handles; they leave this
+ * hook undefined.
+ */
+/**
+ * Resolve `data.Values` / `data.Keys` on a Dictionary-like receiver
+ * to its element-type simple name. Returns `undefined` for any
+ * receiver / accessor combination we don't recognize, letting the
+ * compound-receiver pass fall through to the regular field walk.
+ */
+export declare function unwrapCsharpCollectionAccessor(receiverType: string, accessor: string): string | undefined;

package/dist/core/ingestion/languages/csharp/accessor-unwrap.js

@@ -0,0 +1,56 @@
+/**
+ * C# collection-accessor unwrapping.
+ *
+ * When the compound-receiver resolver encounters a trailing
+ * `.Values` / `.Keys` on a dotted member-access chain, it calls the
+ * provider's `unwrapCollectionAccessor` hook to find the element
+ * type. This module supplies the C# implementation — recognizing
+ * Dictionary-family generics and returning the value or key type.
+ *
+ * Other languages (Python, Java, TypeScript) use method-call syntax
+ * for the same access (`.values()` / `.keys()`), which the compound-
+ * receiver's call-expression branch already handles; they leave this
+ * hook undefined.
+ */
+/** Extract (K, V) from `Dictionary<K, V>` / `IDictionary<K, V>` /
+ *  `IReadOnlyDictionary<K, V>` / `SortedDictionary<K, V>` /
+ *  `ConcurrentDictionary<K, V>` / `ImmutableDictionary<K, V>`.
+ *  Returns undefined if the type name doesn't match or the argument
+ *  list isn't exactly two top-level args. */
+function extractDictionaryArgs(rawName) {
+    const match = rawName.match(/^(?:[A-Za-z_][A-Za-z0-9_.]*\.)?(?:Dictionary|IDictionary|IReadOnlyDictionary|SortedDictionary|ConcurrentDictionary|ImmutableDictionary)<(.+)>$/);
+    if (match === null)
+        return undefined;
+    const inner = match[1];
+    // Split on the top-level comma (tolerate nested `<...>`).
+    let depth = 0;
+    let commaIdx = -1;
+    for (let i = 0; i < inner.length; i++) {
+        const ch = inner[i];
+        if (ch === '<')
+            depth++;
+        else if (ch === '>')
+            depth--;
+        else if (ch === ',' && depth === 0) {
+            commaIdx = i;
+            break;
+        }
+    }
+    if (commaIdx === -1)
+        return undefined;
+    return { key: inner.slice(0, commaIdx).trim(), value: inner.slice(commaIdx + 1).trim() };
+}
+/**
+ * Resolve `data.Values` / `data.Keys` on a Dictionary-like receiver
+ * to its element-type simple name. Returns `undefined` for any
+ * receiver / accessor combination we don't recognize, letting the
+ * compound-receiver pass fall through to the regular field walk.
+ */
+export function unwrapCsharpCollectionAccessor(receiverType, accessor) {
+    if (accessor !== 'Values' && accessor !== 'Keys')
+        return undefined;
+    const args = extractDictionaryArgs(receiverType);
+    if (args === undefined)
+        return undefined;
+    return accessor === 'Values' ? args.value : args.key;
+}

package/dist/core/ingestion/languages/csharp/arity-metadata.d.ts

@@ -0,0 +1,26 @@
+/**
+ * Extract C# arity metadata from a method-like tree-sitter node —
+ * `method_declaration`, `constructor_declaration`, `destructor_declaration`,
+ * `operator_declaration`, `conversion_operator_declaration`, or
+ * `local_function_statement`.
+ *
+ * Reuses `csharpMethodConfig.extractParameters` so scope-extracted defs
+ * carry the same arity semantics as the legacy parse-worker path:
+ *   - `params` variadic collapses `parameterCount` to `undefined`,
+ *     which `csharpArityCompatibility` then treats as "max unknown" —
+ *     the candidate stays eligible at `argCount >= required`.
+ *   - Defaulted parameters (`= expr`) contribute to `optionalCount`;
+ *     `requiredParameterCount = total − optionalCount`.
+ *   - `parameterTypes` collects declared type names (with `ref`/`out`/
+ *     `in` prefix) for overload narrowing; a literal `'params'` marker
+ *     is appended for variadic methods so `csharpArityCompatibility`
+ *     can detect them without re-reading the AST.
+ */
+import type { SyntaxNode } from '../../utils/ast-helpers.js';
+interface CsharpArityMetadata {
+    readonly parameterCount: number | undefined;
+    readonly requiredParameterCount: number | undefined;
+    readonly parameterTypes: readonly string[] | undefined;
+}
+export declare function computeCsharpArityMetadata(fnNode: SyntaxNode): CsharpArityMetadata;
+export {};

package/dist/core/ingestion/languages/csharp/arity-metadata.js

@@ -0,0 +1,46 @@
+/**
+ * Extract C# arity metadata from a method-like tree-sitter node —
+ * `method_declaration`, `constructor_declaration`, `destructor_declaration`,
+ * `operator_declaration`, `conversion_operator_declaration`, or
+ * `local_function_statement`.
+ *
+ * Reuses `csharpMethodConfig.extractParameters` so scope-extracted defs
+ * carry the same arity semantics as the legacy parse-worker path:
+ *   - `params` variadic collapses `parameterCount` to `undefined`,
+ *     which `csharpArityCompatibility` then treats as "max unknown" —
+ *     the candidate stays eligible at `argCount >= required`.
+ *   - Defaulted parameters (`= expr`) contribute to `optionalCount`;
+ *     `requiredParameterCount = total − optionalCount`.
+ *   - `parameterTypes` collects declared type names (with `ref`/`out`/
+ *     `in` prefix) for overload narrowing; a literal `'params'` marker
+ *     is appended for variadic methods so `csharpArityCompatibility`
+ *     can detect them without re-reading the AST.
+ */
+import { csharpMethodConfig } from '../../method-extractors/configs/csharp.js';
+export function computeCsharpArityMetadata(fnNode) {
+    const params = csharpMethodConfig.extractParameters?.(fnNode) ?? [];
+    let hasVariadic = false;
+    let optionalCount = 0;
+    const types = [];
+    for (const p of params) {
+        if (p.isVariadic)
+            hasVariadic = true;
+        else if (p.isOptional)
+            optionalCount++;
+        if (p.type !== null)
+            types.push(p.type);
+    }
+    if (hasVariadic)
+        types.push('params');
+    const total = params.length;
+    // `params int[] args` declares one formal param but accepts any arg
+    // count ≥ required — mirror Python's treatment of `*args` and leave
+    // `parameterCount` undefined so the registry treats max as unknown.
+    const parameterCount = hasVariadic ? undefined : total;
+    const requiredParameterCount = hasVariadic ? undefined : total - optionalCount;
+    return {
+        parameterCount,
+        requiredParameterCount,
+        parameterTypes: types.length > 0 ? types : undefined,
+    };
+}

package/dist/core/ingestion/languages/csharp/arity.d.ts

@@ -0,0 +1,23 @@
+/**
+ * C# arity check, accommodating `params` variadic and default parameters.
+ *
+ * The `def` metadata we care about (synthesized by `arity-metadata.ts`):
+ *   - `parameterCount`         — total formal parameters; `undefined`
+ *                                when the method has `params T[]` variadic.
+ *   - `requiredParameterCount` — min required (excludes defaulted params
+ *                                and `params` variadic).
+ *   - `parameterTypes`         — declared type strings; contains the
+ *                                literal `'params'` when the method is
+ *                                variadic.
+ *
+ * Verdicts:
+ *   - `'compatible'`   — `requiredParameterCount <= argCount <= parameterCount`,
+ *                        OR the def takes `params` (then any `argCount >= required`).
+ *   - `'incompatible'` — argCount is below required, OR above max with no variadic.
+ *   - `'unknown'`      — metadata is absent / incomplete.
+ *
+ * `'incompatible'` is a soft signal in `Registry.lookup` (penalized but
+ * still considered when no compatible candidate exists), per RFC §4.
+ */
+import type { Callsite, SymbolDefinition } from '../../../../_shared/index.js';
+export declare function csharpArityCompatibility(def: SymbolDefinition, callsite: Callsite): 'compatible' | 'unknown' | 'incompatible';

package/dist/core/ingestion/languages/csharp/arity.js

@@ -0,0 +1,37 @@
+/**
+ * C# arity check, accommodating `params` variadic and default parameters.
+ *
+ * The `def` metadata we care about (synthesized by `arity-metadata.ts`):
+ *   - `parameterCount`         — total formal parameters; `undefined`
+ *                                when the method has `params T[]` variadic.
+ *   - `requiredParameterCount` — min required (excludes defaulted params
+ *                                and `params` variadic).
+ *   - `parameterTypes`         — declared type strings; contains the
+ *                                literal `'params'` when the method is
+ *                                variadic.
+ *
+ * Verdicts:
+ *   - `'compatible'`   — `requiredParameterCount <= argCount <= parameterCount`,
+ *                        OR the def takes `params` (then any `argCount >= required`).
+ *   - `'incompatible'` — argCount is below required, OR above max with no variadic.
+ *   - `'unknown'`      — metadata is absent / incomplete.
+ *
+ * `'incompatible'` is a soft signal in `Registry.lookup` (penalized but
+ * still considered when no compatible candidate exists), per RFC §4.
+ */
+export function csharpArityCompatibility(def, callsite) {
+    const max = def.parameterCount;
+    const min = def.requiredParameterCount;
+    if (max === undefined && min === undefined)
+        return 'unknown';
+    const argCount = callsite.arity;
+    if (!Number.isFinite(argCount) || argCount < 0)
+        return 'unknown';
+    const hasVarArgs = def.parameterTypes !== undefined &&
+        def.parameterTypes.some((t) => t === 'params' || t.startsWith('params '));
+    if (min !== undefined && argCount < min)
+        return 'incompatible';
+    if (max !== undefined && argCount > max && !hasVarArgs)
+        return 'incompatible';
+    return 'compatible';
+}

package/dist/core/ingestion/languages/csharp/cache-stats.d.ts

@@ -0,0 +1,15 @@
+/**
+ * Dev-mode counters for the cross-phase scope-captures parse cache
+ * (C# mirror of `languages/python/cache-stats.ts`).
+ *
+ * Gated by `PROF_SCOPE_RESOLUTION=1`. Production builds fold every
+ * increment into dead code via the module-level `PROF` constant, so
+ * the hot path in `captures.ts` stays branch-free.
+ */
+export declare function recordCacheHit(): void;
+export declare function recordCacheMiss(): void;
+export declare function getCsharpCaptureCacheStats(): {
+    hits: number;
+    misses: number;
+};
+export declare function resetCsharpCaptureCacheStats(): void;

package/dist/core/ingestion/languages/csharp/cache-stats.js

@@ -0,0 +1,26 @@
+/**
+ * Dev-mode counters for the cross-phase scope-captures parse cache
+ * (C# mirror of `languages/python/cache-stats.ts`).
+ *
+ * Gated by `PROF_SCOPE_RESOLUTION=1`. Production builds fold every
+ * increment into dead code via the module-level `PROF` constant, so
+ * the hot path in `captures.ts` stays branch-free.
+ */
+const PROF = process.env.PROF_SCOPE_RESOLUTION === '1';
+let CACHE_HITS = 0;
+let CACHE_MISSES = 0;
+export function recordCacheHit() {
+    if (PROF)
+        CACHE_HITS++;
+}
+export function recordCacheMiss() {
+    if (PROF)
+        CACHE_MISSES++;
+}
+export function getCsharpCaptureCacheStats() {
+    return { hits: CACHE_HITS, misses: CACHE_MISSES };
+}
+export function resetCsharpCaptureCacheStats() {
+    CACHE_HITS = 0;
+    CACHE_MISSES = 0;
+}

package/dist/core/ingestion/languages/csharp/captures.d.ts

@@ -0,0 +1,19 @@
+/**
+ * `emitScopeCaptures` for C#.
+ *
+ * Drives the C# scope query against tree-sitter-c-sharp and groups raw
+ * matches into `CaptureMatch[]` for the central extractor. Layers one
+ * synthesized stream on top today:
+ *
+ *   1. **Decomposed using directives** — each `using_directive` is
+ *      re-emitted with `@import.kind/source/name/alias` markers so
+ *      `interpretCsharpImport` can recover the ParsedImport shape
+ *      without re-parsing raw text (see `import-decomposer.ts`).
+ *
+ * Receiver-binding synthesis (`this` / `base` type anchors) and arity
+ * metadata synthesis (Unit 5) layer on top later.
+ *
+ * Pure given the input source text. No I/O, no globals consulted.
+ */
+import type { CaptureMatch } from '../../../../_shared/index.js';
+export declare function emitCsharpScopeCaptures(sourceText: string, _filePath: string, cachedTree?: unknown): readonly CaptureMatch[];