@isaacriehm/cairn-core 0.8.0 → 0.9.2

package/dist/init/curator/regex-prefilter.d.ts

@@ -0,0 +1,54 @@
+/**
+ * Curator pipeline — regex pre-filter (Phase 9a-walker).
+ *
+ * Drops corpus records that the curator subagents would only classify
+ * as noise:
+ *   - test files / fixtures / snapshots
+ *   - generated / build / vendor / migrations
+ *   - .archive/ + .planning/archive/
+ *   - JSX block comments (lots of UI annotation noise)
+ *   - license / SPDX headers
+ *   - JSDoc with only @param/@returns/@see/@throws and < 30 words prose
+ *   - TODO-only or banner-only comments
+ *
+ * Mapper.off_limits_globs is also applied (so the operator's own
+ * exclusion list filters the corpus too).
+ *
+ * Plus a pure-function `stripJsdocTags` that strips `@domain`,
+ * `@orgScope`, `@softDelete`, `@see`, `@param`, `@returns`, `@throws`
+ * scaffolding lines from prose so they don't leak into reducer
+ * output. Validators downstream re-check `jsdoc-tag-leak`; this is
+ * the defense-in-depth at the walker level.
+ *
+ * Drop targets (from curator plan):
+ *   60–80% of the raw corpus is expected to drop here.
+ */
+export type DropReason = "test-file" | "generated-dir" | "archive-dir" | "off-limits-glob" | "jsx-block-comment" | "license-header" | "jsdoc-tag-only" | "todo-or-banner-only" | "below-minimum-prose";
+export interface PrefilterArgs {
+    /** Repo-relative path the block came from. */
+    file: string;
+    /**
+     * Source-kind tag — `comment` records get the JSX + JSDoc-tag-only
+     * filters; `doc` and `rule` records skip those (paragraph + section
+     * shapes are different).
+     */
+    source_kind: "comment" | "doc" | "rule";
+    /** Cleaned prose (after the language-specific marker stripper). */
+    prose: string;
+    /** Raw block (carries surrounding context for the JSX detector). */
+    raw?: string;
+    /** Mapper off-limits globs to honor. Glob matching is prefix/contains-friendly. */
+    offLimitsGlobs?: string[];
+}
+export interface PrefilterResult {
+    drop: boolean;
+    reason?: DropReason;
+    /** Prose with JSDoc scaffolding tags stripped. Caller writes this to corpus.jsonl. */
+    cleanedProse: string;
+}
+export declare function applyPrefilter(args: PrefilterArgs): PrefilterResult;
+/**
+ * Strip JSDoc-style scaffolding tag lines from prose. Defensive layer
+ * before the LLM ever sees them; validators downstream still check.
+ */
+export declare function stripJsdocTags(prose: string): string;

package/dist/init/curator/regex-prefilter.js

@@ -0,0 +1,185 @@
+/**
+ * Curator pipeline — regex pre-filter (Phase 9a-walker).
+ *
+ * Drops corpus records that the curator subagents would only classify
+ * as noise:
+ *   - test files / fixtures / snapshots
+ *   - generated / build / vendor / migrations
+ *   - .archive/ + .planning/archive/
+ *   - JSX block comments (lots of UI annotation noise)
+ *   - license / SPDX headers
+ *   - JSDoc with only @param/@returns/@see/@throws and < 30 words prose
+ *   - TODO-only or banner-only comments
+ *
+ * Mapper.off_limits_globs is also applied (so the operator's own
+ * exclusion list filters the corpus too).
+ *
+ * Plus a pure-function `stripJsdocTags` that strips `@domain`,
+ * `@orgScope`, `@softDelete`, `@see`, `@param`, `@returns`, `@throws`
+ * scaffolding lines from prose so they don't leak into reducer
+ * output. Validators downstream re-check `jsdoc-tag-leak`; this is
+ * the defense-in-depth at the walker level.
+ *
+ * Drop targets (from curator plan):
+ *   60–80% of the raw corpus is expected to drop here.
+ */
+const TEST_FILE_RE = /(?:^|\/)(?:[^/]+\.(?:spec|test)\.(?:[tj]sx?|mjs|cjs)|__tests__\/|e2e\/|fixtures\/|snapshots?\/|__snapshots__\/)/;
+const GENERATED_DIR_RE = /(?:^|\/)(?:migrations|dist|build|generated|vendor|node_modules)\//;
+const ARCHIVE_DIR_RE = /(?:^|\/)\.(?:archive|planning\/archive)\//;
+const TODO_BANNER_RE = /^\s*(?:TODO|FIXME|XXX|HACK|NOTE)\b/;
+const PURE_BANNER_RE = /^[\s\W_]+$/;
+const SPDX_LICENSE_RE = /\b(?:SPDX-License-Identifier|All rights reserved|Licensed under)\b/i;
+const COPYRIGHT_RE = /\bcopyright\b/i;
+const JSDOC_TAG_LINE_RE = /^\s*@(?:domain|orgScope|softDelete|see|param|returns?|throws?|example|deprecated|since|version|author|module|namespace|alias|constant|type|typedef|callback|exports|category|memberof|inheritdoc|override|fileoverview|api|public|private|protected|internal|readonly|abstract|static|access|todo|fires|listens|hideconstructor)\b.*$/gm;
+const MIN_WORDS_AFTER_TAG_STRIP = 12;
+export function applyPrefilter(args) {
+    const cleaned = stripJsdocTags(args.prose);
+    if (TEST_FILE_RE.test(args.file)) {
+        return { drop: true, reason: "test-file", cleanedProse: cleaned };
+    }
+    if (GENERATED_DIR_RE.test(args.file)) {
+        return { drop: true, reason: "generated-dir", cleanedProse: cleaned };
+    }
+    if (ARCHIVE_DIR_RE.test(args.file)) {
+        return { drop: true, reason: "archive-dir", cleanedProse: cleaned };
+    }
+    if (matchesAnyGlob(args.file, args.offLimitsGlobs)) {
+        return { drop: true, reason: "off-limits-glob", cleanedProse: cleaned };
+    }
+    if (args.source_kind === "comment") {
+        if (isJsxBlockComment(args.raw, args.file)) {
+            return { drop: true, reason: "jsx-block-comment", cleanedProse: cleaned };
+        }
+        if (isLicenseHeader(args.raw ?? args.prose)) {
+            return { drop: true, reason: "license-header", cleanedProse: cleaned };
+        }
+        if (isJsdocTagOnly(args.prose, cleaned)) {
+            return { drop: true, reason: "jsdoc-tag-only", cleanedProse: cleaned };
+        }
+    }
+    if (isTodoOrBannerOnly(cleaned)) {
+        return { drop: true, reason: "todo-or-banner-only", cleanedProse: cleaned };
+    }
+    if (countWords(cleaned) < MIN_WORDS_AFTER_TAG_STRIP) {
+        return { drop: true, reason: "below-minimum-prose", cleanedProse: cleaned };
+    }
+    return { drop: false, cleanedProse: cleaned };
+}
+/**
+ * Strip JSDoc-style scaffolding tag lines from prose. Defensive layer
+ * before the LLM ever sees them; validators downstream still check.
+ */
+export function stripJsdocTags(prose) {
+    return prose
+        .replace(JSDOC_TAG_LINE_RE, "")
+        .split("\n")
+        .map((line) => line.trimEnd())
+        .join("\n")
+        .replace(/\n{3,}/g, "\n\n")
+        .trim();
+}
+/**
+ * JSX block comments live inside `{/* … *\/}` — they're UI annotations
+ * and almost never carry decision-bearing prose. Detection: the file
+ * is .tsx/.jsx and either the raw block starts with `{/` (the JSX
+ * wrapper got captured) or the surrounding chars are JSX brackets.
+ *
+ * Conservative — when the source is not .tsx/.jsx we never flag this
+ * reason, since `/* … *\/` in normal .ts is fine.
+ */
+function isJsxBlockComment(raw, file) {
+    if (!/\.(?:[tj])sx$/.test(file))
+        return false;
+    if (raw === undefined)
+        return false;
+    const head = raw.trimStart().slice(0, 4);
+    return head.startsWith("{/*");
+}
+function isLicenseHeader(raw) {
+    const head = raw.slice(0, 1500);
+    return SPDX_LICENSE_RE.test(head) || COPYRIGHT_RE.test(head);
+}
+/**
+ * Returns true when the JSDoc block carries scaffolding tags but
+ * < MIN_WORDS_AFTER_TAG_STRIP words of real prose. Many JSDoc blocks
+ * are pure `@param`/`@returns` lists with a one-line summary —
+ * curator should drop those rather than dispatch a Sonnet call.
+ */
+function isJsdocTagOnly(originalProse, cleaned) {
+    const hadTag = JSDOC_TAG_LINE_RE.test(originalProse);
+    if (!hadTag)
+        return false;
+    return countWords(cleaned) < MIN_WORDS_AFTER_TAG_STRIP;
+}
+function isTodoOrBannerOnly(cleaned) {
+    if (cleaned.length === 0)
+        return true;
+    const lines = cleaned
+        .split("\n")
+        .map((l) => l.trim())
+        .filter((l) => l.length > 0);
+    if (lines.length === 0)
+        return true;
+    let nonTodo = 0;
+    for (const line of lines) {
+        if (TODO_BANNER_RE.test(line))
+            continue;
+        if (PURE_BANNER_RE.test(line))
+            continue;
+        nonTodo += 1;
+    }
+    return nonTodo === 0;
+}
+function countWords(prose) {
+    const stripped = prose.replace(/[^\p{L}\p{N}\s]/gu, " ");
+    const tokens = stripped.split(/\s+/).filter((t) => t.length > 0);
+    return tokens.length;
+}
+// Lightweight glob match — supports `*` and `**` segment wildcards,
+// good enough for `off_limits_globs` patterns like `vendor` or `dist`
+// at any depth. Not a full minimatch; the curator only needs
+// prefix-style matching.
+function matchesAnyGlob(file, globs) {
+    if (globs === undefined || globs.length === 0)
+        return false;
+    for (const g of globs) {
+        if (matchesGlob(file, g))
+            return true;
+    }
+    return false;
+}
+function matchesGlob(file, glob) {
+    const re = globToRegex(glob);
+    return re.test(file);
+}
+function globToRegex(glob) {
+    // Translate ** → `.*`, * → `[^/]*`, escape everything else.
+    let out = "";
+    for (let i = 0; i < glob.length; i++) {
+        const ch = glob[i];
+        if (ch === "*") {
+            if (glob[i + 1] === "*") {
+                out += ".*";
+                i += 1;
+            }
+            else {
+                out += "[^/]*";
+            }
+            continue;
+        }
+        if (ch === "?") {
+            out += "[^/]";
+            continue;
+        }
+        if (ch === undefined)
+            continue;
+        if (".+^$()[]{}|\\".includes(ch)) {
+            out += `\\${ch}`;
+        }
+        else {
+            out += ch;
+        }
+    }
+    return new RegExp(`^${out}$`);
+}
+//# sourceMappingURL=regex-prefilter.js.map

package/dist/init/curator/regex-prefilter.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"regex-prefilter.js","sourceRoot":"","sources":["../../../src/init/curator/regex-prefilter.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;;;;;GAwBG;AAEH,MAAM,YAAY,GAChB,iHAAiH,CAAC;AACpH,MAAM,gBAAgB,GACpB,mEAAmE,CAAC;AACtE,MAAM,cAAc,GAAG,2CAA2C,CAAC;AACnE,MAAM,cAAc,GAAG,oCAAoC,CAAC;AAC5D,MAAM,cAAc,GAAG,YAAY,CAAC;AACpC,MAAM,eAAe,GAAG,qEAAqE,CAAC;AAC9F,MAAM,YAAY,GAAG,gBAAgB,CAAC;AACtC,MAAM,iBAAiB,GACrB,yUAAyU,CAAC;AAa5U,MAAM,yBAAyB,GAAG,EAAE,CAAC;AA0BrC,MAAM,UAAU,cAAc,CAAC,IAAmB;IAChD,MAAM,OAAO,GAAG,cAAc,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC;IAE3C,IAAI,YAAY,CAAC,IAAI,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC;QACjC,OAAO,EAAE,IAAI,EAAE,IAAI,EAAE,MAAM,EAAE,WAAW,EAAE,YAAY,EAAE,OAAO,EAAE,CAAC;IACpE,CAAC;IACD,IAAI,gBAAgB,CAAC,IAAI,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC;QACrC,OAAO,EAAE,IAAI,EAAE,IAAI,EAAE,MAAM,EAAE,eAAe,EAAE,YAAY,EAAE,OAAO,EAAE,CAAC;IACxE,CAAC;IACD,IAAI,cAAc,CAAC,IAAI,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC;QACnC,OAAO,EAAE,IAAI,EAAE,IAAI,EAAE,MAAM,EAAE,aAAa,EAAE,YAAY,EAAE,OAAO,EAAE,CAAC;IACtE,CAAC;IACD,IAAI,cAAc,CAAC,IAAI,CAAC,IAAI,EAAE,IAAI,CAAC,cAAc,CAAC,EAAE,CAAC;QACnD,OAAO,EAAE,IAAI,EAAE,IAAI,EAAE,MAAM,EAAE,iBAAiB,EAAE,YAAY,EAAE,OAAO,EAAE,CAAC;IAC1E,CAAC;IAED,IAAI,IAAI,CAAC,WAAW,KAAK,SAAS,EAAE,CAAC;QACnC,IAAI,iBAAiB,CAAC,IAAI,CAAC,GAAG,EAAE,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC;YAC3C,OAAO,EAAE,IAAI,EAAE,IAAI,EAAE,MAAM,EAAE,mBAAmB,EAAE,YAAY,EAAE,OAAO,EAAE,CAAC;QAC5E,CAAC;QACD,IAAI,eAAe,CAAC,IAAI,CAAC,GAAG,IAAI,IAAI,CAAC,KAAK,CAAC,EAAE,CAAC;YAC5C,OAAO,EAAE,IAAI,EAAE,IAAI,EAAE,MAAM,EAAE,gBAAgB,EAAE,YAAY,EAAE,OAAO,EAAE,CAAC;QACzE,CAAC;QACD,IAAI,cAAc,CAAC,IAAI,CAAC,KAAK,EAAE,OAAO,CAAC,EAAE,CAAC;YACxC,OAAO,EAAE,IAAI,EAAE,IAAI,EAAE,MAAM,EAAE,gBAAgB,EAAE,YAAY,EAAE,OAAO,EAAE,CAAC;QACzE,CAAC;IACH,CAAC;IAED,IAAI,kBAAkB,CAAC,OAAO,CAAC,EAAE,CAAC;QAChC,OAAO,EAAE,IAAI,EAAE,IAAI,EAAE,MAAM,EAAE,qBAAqB,EAAE,YAAY,EAAE,OAAO,EAAE,CAAC;IAC9E,CAAC;IACD,IAAI,UAAU,CAAC,OAAO,CAAC,GAAG,yBAAyB,EAAE,CAAC;QACpD,OAAO,EAAE,IAAI,EAAE,IAAI,EAAE,MAAM,EAAE,qBAAqB,EAAE,YAAY,EAAE,OAAO,EAAE,CAAC;IAC9E,CAAC;IAED,OAAO,EAAE,IAAI,EAAE,KAAK,EAAE,YAAY,EAAE,OAAO,EAAE,CAAC;AAChD,CAAC;AAED;;;GAGG;AACH,MAAM,UAAU,cAAc,CAAC,KAAa;IAC1C,OAAO,KAAK;SACT,OAAO,CAAC,iBAAiB,EAAE,EAAE,CAAC;SAC9B,KAAK,CAAC,IAAI,CAAC;SACX,GAAG,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,CAAC,OAAO,EAAE,CAAC;SAC7B,IAAI,CAAC,IAAI,CAAC;SACV,OAAO,CAAC,SAAS,EAAE,MAAM,CAAC;SAC1B,IAAI,EAAE,CAAC;AACZ,CAAC;AAED;;;;;;;;GAQG;AACH,SAAS,iBAAiB,CAAC,GAAuB,EAAE,IAAY;IAC9D,IAAI,CAAC,eAAe,CAAC,IAAI,CAAC,IAAI,CAAC;QAAE,OAAO,KAAK,CAAC;IAC9C,IAAI,GAAG,KAAK,SAAS;QAAE,OAAO,KAAK,CAAC;IACpC,MAAM,IAAI,GAAG,GAAG,CAAC,SAAS,EAAE,CAAC,KAAK,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC;IACzC,OAAO,IAAI,CAAC,UAAU,CAAC,KAAK,CAAC,CAAC;AAChC,CAAC;AAED,SAAS,eAAe,CAAC,GAAW;IAClC,MAAM,IAAI,GAAG,GAAG,CAAC,KAAK,CAAC,CAAC,EAAE,IAAI,CAAC,CAAC;IAChC,OAAO,eAAe,CAAC,IAAI,CAAC,IAAI,CAAC,IAAI,YAAY,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;AAC/D,CAAC;AAED;;;;;GAKG;AACH,SAAS,cAAc,CAAC,aAAqB,EAAE,OAAe;IAC5D,MAAM,MAAM,GAAG,iBAAiB,CAAC,IAAI,CAAC,aAAa,CAAC,CAAC;IACrD,IAAI,CAAC,MAAM;QAAE,OAAO,KAAK,CAAC;IAC1B,OAAO,UAAU,CAAC,OAAO,CAAC,GAAG,yBAAyB,CAAC;AACzD,CAAC;AAED,SAAS,kBAAkB,CAAC,OAAe;IACzC,IAAI,OAAO,CAAC,MAAM,KAAK,CAAC;QAAE,OAAO,IAAI,CAAC;IACtC,MAAM,KAAK,GAAG,OAAO;SAClB,KAAK,CAAC,IAAI,CAAC;SACX,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC;SACpB,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC;IAC/B,IAAI,KAAK,CAAC,MAAM,KAAK,CAAC;QAAE,OAAO,IAAI,CAAC;IACpC,IAAI,OAAO,GAAG,CAAC,CAAC;IAChB,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;QACzB,IAAI,cAAc,CAAC,IAAI,CAAC,IAAI,CAAC;YAAE,SAAS;QACxC,IAAI,cAAc,CAAC,IAAI,CAAC,IAAI,CAAC;YAAE,SAAS;QACxC,OAAO,IAAI,CAAC,CAAC;IACf,CAAC;IACD,OAAO,OAAO,KAAK,CAAC,CAAC;AACvB,CAAC;AAED,SAAS,UAAU,CAAC,KAAa;IAC/B,MAAM,QAAQ,GAAG,KAAK,CAAC,OAAO,CAAC,mBAAmB,EAAE,GAAG,CAAC,CAAC;IACzD,MAAM,MAAM,GAAG,QAAQ,CAAC,KAAK,CAAC,KAAK,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC;IACjE,OAAO,MAAM,CAAC,MAAM,CAAC;AACvB,CAAC;AAED,oEAAoE;AACpE,sEAAsE;AACtE,6DAA6D;AAC7D,yBAAyB;AACzB,SAAS,cAAc,CAAC,IAAY,EAAE,KAAgB;IACpD,IAAI,KAAK,KAAK,SAAS,IAAI,KAAK,CAAC,MAAM,KAAK,CAAC;QAAE,OAAO,KAAK,CAAC;IAC5D,KAAK,MAAM,CAAC,IAAI,KAAK,EAAE,CAAC;QACtB,IAAI,WAAW,CAAC,IAAI,EAAE,CAAC,CAAC;YAAE,OAAO,IAAI,CAAC;IACxC,CAAC;IACD,OAAO,KAAK,CAAC;AACf,CAAC;AAED,SAAS,WAAW,CAAC,IAAY,EAAE,IAAY;IAC7C,MAAM,EAAE,GAAG,WAAW,CAAC,IAAI,CAAC,CAAC;IAC7B,OAAO,EAAE,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;AACvB,CAAC;AAED,SAAS,WAAW,CAAC,IAAY;IAC/B,4DAA4D;IAC5D,IAAI,GAAG,GAAG,EAAE,CAAC;IACb,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,IAAI,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;QACrC,MAAM,EAAE,GAAG,IAAI,CAAC,CAAC,CAAC,CAAC;QACnB,IAAI,EAAE,KAAK,GAAG,EAAE,CAAC;YACf,IAAI,IAAI,CAAC,CAAC,GAAG,CAAC,CAAC,KAAK,GAAG,EAAE,CAAC;gBACxB,GAAG,IAAI,IAAI,CAAC;gBACZ,CAAC,IAAI,CAAC,CAAC;YACT,CAAC;iBAAM,CAAC;gBACN,GAAG,IAAI,OAAO,CAAC;YACjB,CAAC;YACD,SAAS;QACX,CAAC;QACD,IAAI,EAAE,KAAK,GAAG,EAAE,CAAC;YACf,GAAG,IAAI,MAAM,CAAC;YACd,SAAS;QACX,CAAC;QACD,IAAI,EAAE,KAAK,SAAS;YAAE,SAAS;QAC/B,IAAI,eAAe,CAAC,QAAQ,CAAC,EAAE,CAAC,EAAE,CAAC;YACjC,GAAG,IAAI,KAAK,EAAE,EAAE,CAAC;QACnB,CAAC;aAAM,CAAC;YACN,GAAG,IAAI,EAAE,CAAC;QACZ,CAAC;IACH,CAAC;IACD,OAAO,IAAI,MAAM,CAAC,IAAI,GAAG,GAAG,CAAC,CAAC;AAChC,CAAC"}

package/dist/init/curator/validate.d.ts

@@ -0,0 +1,46 @@
+/**
+ * Curator pipeline — strict per-entry validators.
+ *
+ * Phase 9c-emit feeds every reducer-output entry through `validateEntry`
+ * before writing it to `.cairn/ground/decisions/` or
+ * `.cairn/ground/invariants/`. Operator's auto-accept directive
+ * (curator plan §"Decision log" Q2) requires a high quality bar — when
+ * an entry fails any check it is dropped silently with a counter
+ * logged, never falling back to `_inbox/`.
+ *
+ * Failure modes encoded by `rejectReason`:
+ *   - `title-length`              — empty or > 80 chars
+ *   - `title-no-cap`              — does not start with an uppercase letter
+ *   - `title-trailing-punct`      — ends in `,` `:` `;`
+ *   - `title-truncated-or-jsx`    — ends in `...` or starts with `{/*` (JSX
+ *                                   block-comment leakage)
+ *   - `body-missing-<section>`    — required heading missing
+ *   - `jsdoc-tag-leak`            — body contains `@domain`, `@orgScope`,
+ *                                   `@softDelete`, `@see`, `@param`,
+ *                                   `@returns`, `@throws` (curator pasted
+ *                                   raw scaffolding)
+ *   - `title-pasted-in-body`      — body contains the title verbatim
+ *                                   (indicates unsynthesized pass-through)
+ *   - `no-scope-globs`            — empty `scope_globs`
+ *   - `no-evidence`               — empty `evidence_files`
+ *   - `evidence-missing:<path>`   — cited evidence file does not exist
+ */
+export interface FinalEntry {
+    kind: "DEC" | "INV";
+    title: string;
+    body: string;
+    scope_globs: string[];
+    evidence_files: string[];
+    topic_tags: string[];
+}
+export interface ValidationResult {
+    valid: boolean;
+    rejectReason?: string;
+}
+export declare function validateEntry(e: FinalEntry, repoRoot: string): ValidationResult;
+/**
+ * Strip the `:42-58` or `#L42-L58` line-range suffix off an evidence
+ * file reference. Both forms are accepted on input; the validator
+ * resolves to the bare path before checking file existence.
+ */
+export declare function stripLineRange(evidenceFile: string): string;

package/dist/init/curator/validate.js

@@ -0,0 +1,100 @@
+/**
+ * Curator pipeline — strict per-entry validators.
+ *
+ * Phase 9c-emit feeds every reducer-output entry through `validateEntry`
+ * before writing it to `.cairn/ground/decisions/` or
+ * `.cairn/ground/invariants/`. Operator's auto-accept directive
+ * (curator plan §"Decision log" Q2) requires a high quality bar — when
+ * an entry fails any check it is dropped silently with a counter
+ * logged, never falling back to `_inbox/`.
+ *
+ * Failure modes encoded by `rejectReason`:
+ *   - `title-length`              — empty or > 80 chars
+ *   - `title-no-cap`              — does not start with an uppercase letter
+ *   - `title-trailing-punct`      — ends in `,` `:` `;`
+ *   - `title-truncated-or-jsx`    — ends in `...` or starts with `{/*` (JSX
+ *                                   block-comment leakage)
+ *   - `body-missing-<section>`    — required heading missing
+ *   - `jsdoc-tag-leak`            — body contains `@domain`, `@orgScope`,
+ *                                   `@softDelete`, `@see`, `@param`,
+ *                                   `@returns`, `@throws` (curator pasted
+ *                                   raw scaffolding)
+ *   - `title-pasted-in-body`      — body contains the title verbatim
+ *                                   (indicates unsynthesized pass-through)
+ *   - `no-scope-globs`            — empty `scope_globs`
+ *   - `no-evidence`               — empty `evidence_files`
+ *   - `evidence-missing:<path>`   — cited evidence file does not exist
+ */
+import { existsSync } from "node:fs";
+import { join } from "node:path";
+export function validateEntry(e, repoRoot) {
+    // Title — order matters: catch the most-specific failure modes
+    // (truncation / JSX leakage) BEFORE the generic capitalization check
+    // so a JSX-prefixed title doesn't get reported as merely
+    // `title-no-cap`.
+    if (e.title.length === 0 || e.title.length > 80) {
+        return { valid: false, rejectReason: "title-length" };
+    }
+    if (/\.\.\.$/.test(e.title) || /^\{\/\*/.test(e.title)) {
+        return { valid: false, rejectReason: "title-truncated-or-jsx" };
+    }
+    if (!/^[A-Z]/.test(e.title)) {
+        return { valid: false, rejectReason: "title-no-cap" };
+    }
+    if (/[,:;]$/.test(e.title)) {
+        return { valid: false, rejectReason: "title-trailing-punct" };
+    }
+    // Body sections
+    const requiredSections = e.kind === "INV"
+        ? ["## Context", "## Invariant", "## Why"]
+        : ["## Context", "## Decision", "## Why"];
+    for (const sec of requiredSections) {
+        if (!e.body.includes(sec)) {
+            return { valid: false, rejectReason: `body-missing-${sec}` };
+        }
+    }
+    // No JSDoc tag leak
+    if (/@(domain|orgScope|softDelete|see|param|returns|throws)\b/.test(e.body)) {
+        return { valid: false, rejectReason: "jsdoc-tag-leak" };
+    }
+    // No title-in-body paste
+    if (e.body.includes(e.title)) {
+        return { valid: false, rejectReason: "title-pasted-in-body" };
+    }
+    // scope_globs nonempty
+    if (e.scope_globs.length === 0) {
+        return { valid: false, rejectReason: "no-scope-globs" };
+    }
+    // evidence_files nonempty + exist
+    if (e.evidence_files.length === 0) {
+        return { valid: false, rejectReason: "no-evidence" };
+    }
+    for (const ev of e.evidence_files) {
+        const path = stripLineRange(ev);
+        if (!existsSync(join(repoRoot, path))) {
+            return { valid: false, rejectReason: `evidence-missing:${path}` };
+        }
+    }
+    return { valid: true };
+}
+/**
+ * Strip the `:42-58` or `#L42-L58` line-range suffix off an evidence
+ * file reference. Both forms are accepted on input; the validator
+ * resolves to the bare path before checking file existence.
+ */
+export function stripLineRange(evidenceFile) {
+    // GitHub-style anchor: `path/to/file.ts#L42-L58`
+    const hashIdx = evidenceFile.indexOf("#");
+    if (hashIdx !== -1)
+        return evidenceFile.slice(0, hashIdx);
+    // Colon-style range: `path/to/file.ts:42-58`. Only strip when the
+    // suffix is digit-only (avoids clobbering Windows drive letters
+    // `C:/...` — but those are absolute and should never appear in a
+    // repo-relative evidence path; defensive anyway).
+    const colonIdx = evidenceFile.lastIndexOf(":");
+    if (colonIdx > 1 && /^\d+(?:-\d+)?$/.test(evidenceFile.slice(colonIdx + 1))) {
+        return evidenceFile.slice(0, colonIdx);
+    }
+    return evidenceFile;
+}
+//# sourceMappingURL=validate.js.map

package/dist/init/curator/validate.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"validate.js","sourceRoot":"","sources":["../../../src/init/curator/validate.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;;;;;;;GA0BG;AAEH,OAAO,EAAE,UAAU,EAAE,MAAM,SAAS,CAAC;AACrC,OAAO,EAAE,IAAI,EAAE,MAAM,WAAW,CAAC;AAgBjC,MAAM,UAAU,aAAa,CAC3B,CAAa,EACb,QAAgB;IAEhB,+DAA+D;IAC/D,qEAAqE;IACrE,yDAAyD;IACzD,kBAAkB;IAClB,IAAI,CAAC,CAAC,KAAK,CAAC,MAAM,KAAK,CAAC,IAAI,CAAC,CAAC,KAAK,CAAC,MAAM,GAAG,EAAE,EAAE,CAAC;QAChD,OAAO,EAAE,KAAK,EAAE,KAAK,EAAE,YAAY,EAAE,cAAc,EAAE,CAAC;IACxD,CAAC;IACD,IAAI,SAAS,CAAC,IAAI,CAAC,CAAC,CAAC,KAAK,CAAC,IAAI,SAAS,CAAC,IAAI,CAAC,CAAC,CAAC,KAAK,CAAC,EAAE,CAAC;QACvD,OAAO,EAAE,KAAK,EAAE,KAAK,EAAE,YAAY,EAAE,wBAAwB,EAAE,CAAC;IAClE,CAAC;IACD,IAAI,CAAC,QAAQ,CAAC,IAAI,CAAC,CAAC,CAAC,KAAK,CAAC,EAAE,CAAC;QAC5B,OAAO,EAAE,KAAK,EAAE,KAAK,EAAE,YAAY,EAAE,cAAc,EAAE,CAAC;IACxD,CAAC;IACD,IAAI,QAAQ,CAAC,IAAI,CAAC,CAAC,CAAC,KAAK,CAAC,EAAE,CAAC;QAC3B,OAAO,EAAE,KAAK,EAAE,KAAK,EAAE,YAAY,EAAE,sBAAsB,EAAE,CAAC;IAChE,CAAC;IAED,gBAAgB;IAChB,MAAM,gBAAgB,GAAG,CAAC,CAAC,IAAI,KAAK,KAAK;QACvC,CAAC,CAAC,CAAC,YAAY,EAAE,cAAc,EAAE,QAAQ,CAAC;QAC1C,CAAC,CAAC,CAAC,YAAY,EAAE,aAAa,EAAE,QAAQ,CAAC,CAAC;IAC5C,KAAK,MAAM,GAAG,IAAI,gBAAgB,EAAE,CAAC;QACnC,IAAI,CAAC,CAAC,CAAC,IAAI,CAAC,QAAQ,CAAC,GAAG,CAAC,EAAE,CAAC;YAC1B,OAAO,EAAE,KAAK,EAAE,KAAK,EAAE,YAAY,EAAE,gBAAgB,GAAG,EAAE,EAAE,CAAC;QAC/D,CAAC;IACH,CAAC;IAED,oBAAoB;IACpB,IAAI,0DAA0D,CAAC,IAAI,CAAC,CAAC,CAAC,IAAI,CAAC,EAAE,CAAC;QAC5E,OAAO,EAAE,KAAK,EAAE,KAAK,EAAE,YAAY,EAAE,gBAAgB,EAAE,CAAC;IAC1D,CAAC;IAED,yBAAyB;IACzB,IAAI,CAAC,CAAC,IAAI,CAAC,QAAQ,CAAC,CAAC,CAAC,KAAK,CAAC,EAAE,CAAC;QAC7B,OAAO,EAAE,KAAK,EAAE,KAAK,EAAE,YAAY,EAAE,sBAAsB,EAAE,CAAC;IAChE,CAAC;IAED,uBAAuB;IACvB,IAAI,CAAC,CAAC,WAAW,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;QAC/B,OAAO,EAAE,KAAK,EAAE,KAAK,EAAE,YAAY,EAAE,gBAAgB,EAAE,CAAC;IAC1D,CAAC;IAED,kCAAkC;IAClC,IAAI,CAAC,CAAC,cAAc,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;QAClC,OAAO,EAAE,KAAK,EAAE,KAAK,EAAE,YAAY,EAAE,aAAa,EAAE,CAAC;IACvD,CAAC;IACD,KAAK,MAAM,EAAE,IAAI,CAAC,CAAC,cAAc,EAAE,CAAC;QAClC,MAAM,IAAI,GAAG,cAAc,CAAC,EAAE,CAAC,CAAC;QAChC,IAAI,CAAC,UAAU,CAAC,IAAI,CAAC,QAAQ,EAAE,IAAI,CAAC,CAAC,EAAE,CAAC;YACtC,OAAO,EAAE,KAAK,EAAE,KAAK,EAAE,YAAY,EAAE,oBAAoB,IAAI,EAAE,EAAE,CAAC;QACpE,CAAC;IACH,CAAC;IAED,OAAO,EAAE,KAAK,EAAE,IAAI,EAAE,CAAC;AACzB,CAAC;AAED;;;;GAIG;AACH,MAAM,UAAU,cAAc,CAAC,YAAoB;IACjD,iDAAiD;IACjD,MAAM,OAAO,GAAG,YAAY,CAAC,OAAO,CAAC,GAAG,CAAC,CAAC;IAC1C,IAAI,OAAO,KAAK,CAAC,CAAC;QAAE,OAAO,YAAY,CAAC,KAAK,CAAC,CAAC,EAAE,OAAO,CAAC,CAAC;IAC1D,kEAAkE;IAClE,gEAAgE;IAChE,iEAAiE;IACjE,kDAAkD;IAClD,MAAM,QAAQ,GAAG,YAAY,CAAC,WAAW,CAAC,GAAG,CAAC,CAAC;IAC/C,IAAI,QAAQ,GAAG,CAAC,IAAI,gBAAgB,CAAC,IAAI,CAAC,YAAY,CAAC,KAAK,CAAC,QAAQ,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC;QAC5E,OAAO,YAAY,CAAC,KAAK,CAAC,CAAC,EAAE,QAAQ,CAAC,CAAC;IACzC,CAAC;IACD,OAAO,YAAY,CAAC;AACtB,CAAC"}

package/dist/init/curator/walker.d.ts

@@ -0,0 +1,36 @@
+/**
+ * Curator pipeline — Phase 9a-walker top-level (v0.9.0).
+ *
+ * Builds the unified corpus by running three sub-walkers:
+ *
+ *   - Source comments (existing `walkSourceComments`) — essay-class
+ *     block comments per source file
+ *   - Doc paragraphs (existing `discoverDocs` + paragraph splitter)
+ *     — README + docs/**\/*.md paragraphs ≥80 chars
+ *   - Rule sections (existing `discoverRuleSources` +
+ *     `parseRuleSections`) — H2/H3 sections from CLAUDE.md /
+ *     AGENTS.md / .claude/rules/**\/*.md
+ *
+ * Each candidate runs through the regex pre-filter (`regex-prefilter.ts`)
+ * which drops 60-80% of raw blocks (test files, JSX comments, license
+ * headers, JSDoc with only @tags, etc.). Survivors get written to
+ * `.cairn/init/curator/corpus.jsonl` and packed into shards capped
+ * at 120k input tokens (`shards.json`).
+ */
+export interface RunCuratorWalkerArgs {
+    repoRoot: string;
+}
+export interface RunCuratorWalkerResult {
+    corpus_path: string;
+    shards_path: string;
+    records_total: number;
+    records_by_kind: {
+        comment: number;
+        doc: number;
+        rule: number;
+    };
+    dropped: Record<string, number>;
+    shards: number;
+    total_input_tokens_estimate: number;
+}
+export declare function runCuratorWalker(args: RunCuratorWalkerArgs): Promise<RunCuratorWalkerResult>;