memkin 0.4.1 → 0.4.3

package/dist/distiller/map-reduce.js

@@ -0,0 +1,92 @@
+/**
+ * Sequential map-reduce distillation (spec §5.1).
+ *
+ * MAP (sequential, not parallel): each segment is distilled in order; the prompt
+ * for segment N carries forward segment N-1's `carry_forward` (in-progress topics
+ * / undecided items) so `overturned` can be filled reliably. Session counts are
+ * low, so sequential cost is acceptable.
+ *
+ * Intermediate per-segment schema:
+ *   { seg_no, summary, tentative_signals[], overturned[], carry_forward }
+ *
+ * REDUCE: merges the tentative signals into a final payload. The reduce prompt
+ * carries the explicit rule "conclusions appearing in any `overturned` must not
+ * enter the final signals". We additionally post-filter overturned topics
+ * programmatically so a model slip cannot leak a retracted conclusion back in.
+ *
+ * The final payload is only structurally validated here (parsePayload). Evidence
+ * bounds + reference.url locatability (msg-id.ts) and second-pass privacy run in
+ * the orchestrator, which has the full transcript.
+ */
+import { z } from "zod";
+import { DistilledPayloadSchema, parsePayload, slugifyTopic, } from "./contract.js";
+import { getMapPrompt, getReducePrompt } from "./prompt.js";
+// Intermediate map output. tentative_signals / overturned stay loose (z.unknown /
+// partial) because a segment view is provisional — only the reduced payload is
+// held to the strict contract.
+export const OverturnedSchema = z.object({
+    topic: z.string().min(1),
+    reason: z.string().optional(),
+});
+export const SegmentSummarySchema = z.object({
+    seg_no: z.number(),
+    summary: z.string(),
+    tentative_signals: z.array(z.record(z.unknown())),
+    overturned: z.array(OverturnedSchema),
+    carry_forward: z.string(),
+});
+function renderSegment(seg) {
+    return seg.messages
+        .map((m) => {
+        const cont = m.continued ? " (continued)" : "";
+        return `[${m.msgId}${cont}] ${m.role}: ${m.content}`;
+    })
+        .join("\n");
+}
+/**
+ * Run sequential map over segments then a single reduce.
+ * `criteria` is the judgment-declaration prompt fragment (spec §5 / §11 task 2.9);
+ * callers inject the embedded distill criteria.
+ */
+export async function mapReduceDistill(segments, provider, criteria) {
+    const summaries = [];
+    let carry = "";
+    for (const seg of segments) {
+        const prompt = getMapPrompt({
+            criteria,
+            segmentText: renderSegment(seg),
+            segNo: seg.segNo,
+            carryForward: carry,
+        });
+        const raw = await provider.chat([
+            { role: "system", content: "You are a session distiller. Reply with JSON only." },
+            { role: "user", content: prompt },
+        ], { responseFormat: "json" });
+        const parsed = SegmentSummarySchema.parse(JSON.parse(raw));
+        summaries.push(parsed);
+        carry = parsed.carry_forward;
+    }
+    const overturnedTopics = summaries.flatMap((s) => s.overturned.map((o) => o.topic));
+    const overturnedSlugs = new Set(overturnedTopics.map(slugifyTopic));
+    const reducePrompt = getReducePrompt({
+        criteria,
+        segmentSummaries: summaries,
+        overturnedTopics,
+    });
+    const reduceRaw = await provider.chat([
+        { role: "system", content: "You are a session distiller. Reply with JSON only." },
+        { role: "user", content: reducePrompt },
+    ], { responseFormat: "json" });
+    const reduceParsed = DistilledPayloadSchema.parse(JSON.parse(reduceRaw));
+    // Belt-and-suspenders: strip any signal whose topic was overturned, even if the
+    // reducer wrongly re-included it.
+    const filtered = {
+        signals: reduceParsed.signals.filter((s) => !overturnedSlugs.has(slugifyTopic(s.topic))),
+    };
+    const finalRes = parsePayload(filtered);
+    if (!finalRes.ok) {
+        throw new Error(`reduced payload failed validation: ${finalRes.error.message}`);
+    }
+    return { payload: finalRes.payload, overturnedTopics, segmentSummaries: summaries };
+}
+//# sourceMappingURL=map-reduce.js.map

package/dist/distiller/map-reduce.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"map-reduce.js","sourceRoot":"","sources":["../../src/distiller/map-reduce.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;GAmBG;AAEH,OAAO,EAAE,CAAC,EAAE,MAAM,KAAK,CAAC;AAExB,OAAO,EAEL,sBAAsB,EACtB,YAAY,EACZ,YAAY,GACb,MAAM,eAAe,CAAC;AACvB,OAAO,EAAE,YAAY,EAAE,eAAe,EAAE,MAAM,aAAa,CAAC;AAG5D,kFAAkF;AAClF,+EAA+E;AAC/E,+BAA+B;AAC/B,MAAM,CAAC,MAAM,gBAAgB,GAAG,CAAC,CAAC,MAAM,CAAC;IACvC,KAAK,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,GAAG,CAAC,CAAC,CAAC;IACxB,MAAM,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,QAAQ,EAAE;CAC9B,CAAC,CAAC;AAEH,MAAM,CAAC,MAAM,oBAAoB,GAAG,CAAC,CAAC,MAAM,CAAC;IAC3C,MAAM,EAAE,CAAC,CAAC,MAAM,EAAE;IAClB,OAAO,EAAE,CAAC,CAAC,MAAM,EAAE;IACnB,iBAAiB,EAAE,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,OAAO,EAAE,CAAC,CAAC;IACjD,UAAU,EAAE,CAAC,CAAC,KAAK,CAAC,gBAAgB,CAAC;IACrC,aAAa,EAAE,CAAC,CAAC,MAAM,EAAE;CAC1B,CAAC,CAAC;AASH,SAAS,aAAa,CAAC,GAAY;IACjC,OAAO,GAAG,CAAC,QAAQ;SAChB,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE;QACT,MAAM,IAAI,GAAG,CAAC,CAAC,SAAS,CAAC,CAAC,CAAC,cAAc,CAAC,CAAC,CAAC,EAAE,CAAC;QAC/C,OAAO,IAAI,CAAC,CAAC,KAAK,GAAG,IAAI,KAAK,CAAC,CAAC,IAAI,KAAK,CAAC,CAAC,OAAO,EAAE,CAAC;IACvD,CAAC,CAAC;SACD,IAAI,CAAC,IAAI,CAAC,CAAC;AAChB,CAAC;AAED;;;;GAIG;AACH,MAAM,CAAC,KAAK,UAAU,gBAAgB,CACpC,QAAmB,EACnB,QAAqB,EACrB,QAAiB;IAEjB,MAAM,SAAS,GAAqB,EAAE,CAAC;IACvC,IAAI,KAAK,GAAG,EAAE,CAAC;IAEf,KAAK,MAAM,GAAG,IAAI,QAAQ,EAAE,CAAC;QAC3B,MAAM,MAAM,GAAG,YAAY,CAAC;YAC1B,QAAQ;YACR,WAAW,EAAE,aAAa,CAAC,GAAG,CAAC;YAC/B,KAAK,EAAE,GAAG,CAAC,KAAK;YAChB,YAAY,EAAE,KAAK;SACpB,CAAC,CAAC;QACH,MAAM,GAAG,GAAG,MAAM,QAAQ,CAAC,IAAI,CAC7B;YACE,EAAE,IAAI,EAAE,QAAQ,EAAE,OAAO,EAAE,oDAAoD,EAAE;YACjF,EAAE,IAAI,EAAE,MAAM,EAAE,OAAO,EAAE,MAAM,EAAE;SAClC,EACD,EAAE,cAAc,EAAE,MAAM,EAAE,CAC3B,CAAC;QACF,MAAM,MAAM,GAAG,oBAAoB,CAAC,KAAK,CAAC,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC,CAAC;QAC3D,SAAS,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC;QACvB,KAAK,GAAG,MAAM,CAAC,aAAa,CAAC;IAC/B,CAAC;IAED,MAAM,gBAAgB,GAAG,SAAS,CAAC,OAAO,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,UAAU,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC;IACpF,MAAM,eAAe,GAAG,IAAI,GAAG,CAAC,gBAAgB,CAAC,GAAG,CAAC,YAAY,CAAC,CAAC,CAAC;IAEpE,MAAM,YAAY,GAAG,eAAe,CAAC;QACnC,QAAQ;QACR,gBAAgB,EAAE,SAAS;QAC3B,gBAAgB;KACjB,CAAC,CAAC;IACH,MAAM,SAAS,GAAG,MAAM,QAAQ,CAAC,IAAI,CACnC;QACE,EAAE,IAAI,EAAE,QAAQ,EAAE,OAAO,EAAE,oDAAoD,EAAE;QACjF,EAAE,IAAI,EAAE,MAAM,EAAE,OAAO,EAAE,YAAY,EAAE;KACxC,EACD,EAAE,cAAc,EAAE,MAAM,EAAE,CAC3B,CAAC;IAEF,MAAM,YAAY,GAAG,sBAAsB,CAAC,KAAK,CAAC,IAAI,CAAC,KAAK,CAAC,SAAS,CAAC,CAAC,CAAC;IAEzE,gFAAgF;IAChF,kCAAkC;IAClC,MAAM,QAAQ,GAAG;QACf,OAAO,EAAE,YAAY,CAAC,OAAO,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,eAAe,CAAC,GAAG,CAAC,YAAY,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC;KACzF,CAAC;IACF,MAAM,QAAQ,GAAG,YAAY,CAAC,QAAQ,CAAC,CAAC;IACxC,IAAI,CAAC,QAAQ,CAAC,EAAE,EAAE,CAAC;QACjB,MAAM,IAAI,KAAK,CAAC,sCAAsC,QAAQ,CAAC,KAAK,CAAC,OAAO,EAAE,CAAC,CAAC;IAClF,CAAC;IAED,OAAO,EAAE,OAAO,EAAE,QAAQ,CAAC,OAAO,EAAE,gBAAgB,EAAE,gBAAgB,EAAE,SAAS,EAAE,CAAC;AACtF,CAAC"}

package/dist/distiller/msg-id.d.ts

@@ -0,0 +1,55 @@
+/**
+ * Programmatic msg_id assignment + evidence validation (spec §5.2).
+ *
+ * The pipeline — not the model — assigns each transcript message a sequential
+ * `msg_id` (`msg-1`, `msg-2`, …). Oversized messages get sub-segment ids with a
+ * dotted suffix (`msg-42.1`, `msg-42.2`; see segmenter.ts). The model may only
+ * cite these given ids, and the validation layer here rejects any evidence range
+ * that falls outside the assigned set — killing line-number / id hallucination.
+ *
+ * It also enforces the spec §5 rule that `reference.url` must be locatable inside
+ * the text of its own evidence range: we pull the evidence text and check the url
+ * appears as a substring, so the model cannot conjure a url out of thin air.
+ *
+ * Ordering note: ranges are resolved by INDEX into the ordered message list, not
+ * by parsing the numeric id. This makes plain ids and dotted sub-segment ids
+ * (`msg-42.1`) order uniformly without a bespoke comparator.
+ */
+import type { DistilledSignal } from "./contract.js";
+export type MessageRole = "user" | "assistant";
+export interface RawInputMessage {
+    role: MessageRole;
+    content: string;
+}
+export interface ParsedMessage {
+    msgId: string;
+    role: MessageRole;
+    content: string;
+}
+/** Assign sequential `msg-N` ids to messages in order. */
+export declare function assignMsgIds(messages: RawInputMessage[]): ParsedMessage[];
+interface EvidenceRange {
+    start: string;
+    end: string;
+}
+/**
+ * Concatenate the text of every message in [start, end] (inclusive), resolved by
+ * index. Throws if either bound is unknown or the range is inverted — callers
+ * that want a boolean should go through validateEvidence.
+ */
+export declare function collectEvidenceText(parsed: ParsedMessage[], range: EvidenceRange): string;
+export type ValidateEvidenceResult = {
+    ok: true;
+} | {
+    ok: false;
+    errors: string[];
+};
+/**
+ * Validate every signal's evidence against the assigned msg_id set:
+ *  - both bounds must reference an assigned id;
+ *  - the range must not be inverted;
+ *  - for `reference` signals, the url must appear in the evidence text.
+ */
+export declare function validateEvidence(signals: DistilledSignal[], parsed: ParsedMessage[]): ValidateEvidenceResult;
+export {};
+//# sourceMappingURL=msg-id.d.ts.map

package/dist/distiller/msg-id.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"msg-id.d.ts","sourceRoot":"","sources":["../../src/distiller/msg-id.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;GAgBG;AAEH,OAAO,KAAK,EAAE,eAAe,EAAE,MAAM,eAAe,CAAC;AAErD,MAAM,MAAM,WAAW,GAAG,MAAM,GAAG,WAAW,CAAC;AAE/C,MAAM,WAAW,eAAe;IAC9B,IAAI,EAAE,WAAW,CAAC;IAClB,OAAO,EAAE,MAAM,CAAC;CACjB;AAED,MAAM,WAAW,aAAa;IAC5B,KAAK,EAAE,MAAM,CAAC;IACd,IAAI,EAAE,WAAW,CAAC;IAClB,OAAO,EAAE,MAAM,CAAC;CACjB;AAED,0DAA0D;AAC1D,wBAAgB,YAAY,CAAC,QAAQ,EAAE,eAAe,EAAE,GAAG,aAAa,EAAE,CAMzE;AAED,UAAU,aAAa;IACrB,KAAK,EAAE,MAAM,CAAC;IACd,GAAG,EAAE,MAAM,CAAC;CACb;AAMD;;;;GAIG;AACH,wBAAgB,mBAAmB,CAAC,MAAM,EAAE,aAAa,EAAE,EAAE,KAAK,EAAE,aAAa,GAAG,MAAM,CAUzF;AAED,MAAM,MAAM,sBAAsB,GAAG;IAAE,EAAE,EAAE,IAAI,CAAA;CAAE,GAAG;IAAE,EAAE,EAAE,KAAK,CAAC;IAAC,MAAM,EAAE,MAAM,EAAE,CAAA;CAAE,CAAC;AAEpF;;;;;GAKG;AACH,wBAAgB,gBAAgB,CAC9B,OAAO,EAAE,eAAe,EAAE,EAC1B,MAAM,EAAE,aAAa,EAAE,GACtB,sBAAsB,CAwCxB"}

package/dist/distiller/msg-id.js

@@ -0,0 +1,91 @@
+/**
+ * Programmatic msg_id assignment + evidence validation (spec §5.2).
+ *
+ * The pipeline — not the model — assigns each transcript message a sequential
+ * `msg_id` (`msg-1`, `msg-2`, …). Oversized messages get sub-segment ids with a
+ * dotted suffix (`msg-42.1`, `msg-42.2`; see segmenter.ts). The model may only
+ * cite these given ids, and the validation layer here rejects any evidence range
+ * that falls outside the assigned set — killing line-number / id hallucination.
+ *
+ * It also enforces the spec §5 rule that `reference.url` must be locatable inside
+ * the text of its own evidence range: we pull the evidence text and check the url
+ * appears as a substring, so the model cannot conjure a url out of thin air.
+ *
+ * Ordering note: ranges are resolved by INDEX into the ordered message list, not
+ * by parsing the numeric id. This makes plain ids and dotted sub-segment ids
+ * (`msg-42.1`) order uniformly without a bespoke comparator.
+ */
+/** Assign sequential `msg-N` ids to messages in order. */
+export function assignMsgIds(messages) {
+    return messages.map((m, i) => ({
+        msgId: `msg-${i + 1}`,
+        role: m.role,
+        content: m.content,
+    }));
+}
+function indexOfMsgId(parsed, msgId) {
+    return parsed.findIndex((m) => m.msgId === msgId);
+}
+/**
+ * Concatenate the text of every message in [start, end] (inclusive), resolved by
+ * index. Throws if either bound is unknown or the range is inverted — callers
+ * that want a boolean should go through validateEvidence.
+ */
+export function collectEvidenceText(parsed, range) {
+    const s = indexOfMsgId(parsed, range.start);
+    const e = indexOfMsgId(parsed, range.end);
+    if (s < 0)
+        throw new Error(`unknown evidence start id: ${range.start}`);
+    if (e < 0)
+        throw new Error(`unknown evidence end id: ${range.end}`);
+    if (s > e)
+        throw new Error(`inverted evidence range: ${range.start}..${range.end}`);
+    return parsed
+        .slice(s, e + 1)
+        .map((m) => m.content)
+        .join("\n");
+}
+/**
+ * Validate every signal's evidence against the assigned msg_id set:
+ *  - both bounds must reference an assigned id;
+ *  - the range must not be inverted;
+ *  - for `reference` signals, the url must appear in the evidence text.
+ */
+export function validateEvidence(signals, parsed) {
+    const errors = [];
+    for (const sig of signals) {
+        for (const range of sig.evidence) {
+            const s = indexOfMsgId(parsed, range.start);
+            const e = indexOfMsgId(parsed, range.end);
+            if (s < 0) {
+                errors.push(`signal "${sig.topic}": evidence start "${range.start}" is out of bounds`);
+                continue;
+            }
+            if (e < 0) {
+                errors.push(`signal "${sig.topic}": evidence end "${range.end}" is out of bounds`);
+                continue;
+            }
+            if (s > e) {
+                errors.push(`signal "${sig.topic}": inverted evidence range ${range.start}..${range.end}`);
+            }
+        }
+        if (sig.type === "reference") {
+            const found = sig.evidence.some((range) => {
+                const s = indexOfMsgId(parsed, range.start);
+                const e = indexOfMsgId(parsed, range.end);
+                if (s < 0 || e < 0 || s > e)
+                    return false;
+                const text = parsed
+                    .slice(s, e + 1)
+                    .map((m) => m.content)
+                    .join("\n");
+                return text.includes(sig.url);
+            });
+            if (!found) {
+                errors.push(`signal "${sig.topic}": reference.url "${sig.url}" not found in its evidence text (possible hallucination)`);
+            }
+        }
+    }
+    return errors.length === 0 ? { ok: true } : { ok: false, errors };
+}
+//# sourceMappingURL=msg-id.js.map

package/dist/distiller/msg-id.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"msg-id.js","sourceRoot":"","sources":["../../src/distiller/msg-id.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;GAgBG;AAiBH,0DAA0D;AAC1D,MAAM,UAAU,YAAY,CAAC,QAA2B;IACtD,OAAO,QAAQ,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC;QAC7B,KAAK,EAAE,OAAO,CAAC,GAAG,CAAC,EAAE;QACrB,IAAI,EAAE,CAAC,CAAC,IAAI;QACZ,OAAO,EAAE,CAAC,CAAC,OAAO;KACnB,CAAC,CAAC,CAAC;AACN,CAAC;AAOD,SAAS,YAAY,CAAC,MAAuB,EAAE,KAAa;IAC1D,OAAO,MAAM,CAAC,SAAS,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,KAAK,KAAK,KAAK,CAAC,CAAC;AACpD,CAAC;AAED;;;;GAIG;AACH,MAAM,UAAU,mBAAmB,CAAC,MAAuB,EAAE,KAAoB;IAC/E,MAAM,CAAC,GAAG,YAAY,CAAC,MAAM,EAAE,KAAK,CAAC,KAAK,CAAC,CAAC;IAC5C,MAAM,CAAC,GAAG,YAAY,CAAC,MAAM,EAAE,KAAK,CAAC,GAAG,CAAC,CAAC;IAC1C,IAAI,CAAC,GAAG,CAAC;QAAE,MAAM,IAAI,KAAK,CAAC,8BAA8B,KAAK,CAAC,KAAK,EAAE,CAAC,CAAC;IACxE,IAAI,CAAC,GAAG,CAAC;QAAE,MAAM,IAAI,KAAK,CAAC,4BAA4B,KAAK,CAAC,GAAG,EAAE,CAAC,CAAC;IACpE,IAAI,CAAC,GAAG,CAAC;QAAE,MAAM,IAAI,KAAK,CAAC,4BAA4B,KAAK,CAAC,KAAK,KAAK,KAAK,CAAC,GAAG,EAAE,CAAC,CAAC;IACpF,OAAO,MAAM;SACV,KAAK,CAAC,CAAC,EAAE,CAAC,GAAG,CAAC,CAAC;SACf,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,OAAO,CAAC;SACrB,IAAI,CAAC,IAAI,CAAC,CAAC;AAChB,CAAC;AAID;;;;;GAKG;AACH,MAAM,UAAU,gBAAgB,CAC9B,OAA0B,EAC1B,MAAuB;IAEvB,MAAM,MAAM,GAAa,EAAE,CAAC;IAE5B,KAAK,MAAM,GAAG,IAAI,OAAO,EAAE,CAAC;QAC1B,KAAK,MAAM,KAAK,IAAI,GAAG,CAAC,QAAQ,EAAE,CAAC;YACjC,MAAM,CAAC,GAAG,YAAY,CAAC,MAAM,EAAE,KAAK,CAAC,KAAK,CAAC,CAAC;YAC5C,MAAM,CAAC,GAAG,YAAY,CAAC,MAAM,EAAE,KAAK,CAAC,GAAG,CAAC,CAAC;YAC1C,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC;gBACV,MAAM,CAAC,IAAI,CAAC,WAAW,GAAG,CAAC,KAAK,sBAAsB,KAAK,CAAC,KAAK,oBAAoB,CAAC,CAAC;gBACvF,SAAS;YACX,CAAC;YACD,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC;gBACV,MAAM,CAAC,IAAI,CAAC,WAAW,GAAG,CAAC,KAAK,oBAAoB,KAAK,CAAC,GAAG,oBAAoB,CAAC,CAAC;gBACnF,SAAS;YACX,CAAC;YACD,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC;gBACV,MAAM,CAAC,IAAI,CAAC,WAAW,GAAG,CAAC,KAAK,8BAA8B,KAAK,CAAC,KAAK,KAAK,KAAK,CAAC,GAAG,EAAE,CAAC,CAAC;YAC7F,CAAC;QACH,CAAC;QAED,IAAI,GAAG,CAAC,IAAI,KAAK,WAAW,EAAE,CAAC;YAC7B,MAAM,KAAK,GAAG,GAAG,CAAC,QAAQ,CAAC,IAAI,CAAC,CAAC,KAAK,EAAE,EAAE;gBACxC,MAAM,CAAC,GAAG,YAAY,CAAC,MAAM,EAAE,KAAK,CAAC,KAAK,CAAC,CAAC;gBAC5C,MAAM,CAAC,GAAG,YAAY,CAAC,MAAM,EAAE,KAAK,CAAC,GAAG,CAAC,CAAC;gBAC1C,IAAI,CAAC,GAAG,CAAC,IAAI,CAAC,GAAG,CAAC,IAAI,CAAC,GAAG,CAAC;oBAAE,OAAO,KAAK,CAAC;gBAC1C,MAAM,IAAI,GAAG,MAAM;qBAChB,KAAK,CAAC,CAAC,EAAE,CAAC,GAAG,CAAC,CAAC;qBACf,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,OAAO,CAAC;qBACrB,IAAI,CAAC,IAAI,CAAC,CAAC;gBACd,OAAO,IAAI,CAAC,QAAQ,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC;YAChC,CAAC,CAAC,CAAC;YACH,IAAI,CAAC,KAAK,EAAE,CAAC;gBACX,MAAM,CAAC,IAAI,CACT,WAAW,GAAG,CAAC,KAAK,qBAAqB,GAAG,CAAC,GAAG,2DAA2D,CAC5G,CAAC;YACJ,CAAC;QACH,CAAC;IACH,CAAC;IAED,OAAO,MAAM,CAAC,MAAM,KAAK,CAAC,CAAC,CAAC,CAAC,EAAE,EAAE,EAAE,IAAI,EAAE,CAAC,CAAC,CAAC,EAAE,EAAE,EAAE,KAAK,EAAE,MAAM,EAAE,CAAC;AACpE,CAAC"}

package/dist/distiller/payload-privacy.d.ts

@@ -0,0 +1,18 @@
+/**
+ * Second-pass privacy over the structured distilled payload (spec §4.3).
+ *
+ * The first pass (RawMessagePrivacyProcessor) redacts raw message text before
+ * the LLM sees it; this second pass runs right before the payload is written to
+ * the outbox, catching sensitive strings the model might have reproduced or
+ * synthesized in its OUTPUT (e.g. reassembling a phone number from context).
+ *
+ * Field policy mirrors PrivacyProcessor (privacy.ts): free-text narrative fields
+ * are redacted; structural fields (type, authority, status, evidence, dates,
+ * enums), entity NAMES (aligned with Entity.name never being redacted) and
+ * reference urls (the url is the core field, checked against evidence) are left
+ * intact.
+ */
+import type { PrivacyConfig } from "../core/config.js";
+import type { DistilledPayload } from "./contract.js";
+export declare function redactPayload(payload: DistilledPayload, config: PrivacyConfig): DistilledPayload;
+//# sourceMappingURL=payload-privacy.d.ts.map

package/dist/distiller/payload-privacy.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"payload-privacy.d.ts","sourceRoot":"","sources":["../../src/distiller/payload-privacy.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;GAaG;AAEH,OAAO,KAAK,EAAE,aAAa,EAAE,MAAM,mBAAmB,CAAC;AACvD,OAAO,KAAK,EAAE,gBAAgB,EAAmB,MAAM,eAAe,CAAC;AAGvE,wBAAgB,aAAa,CAAC,OAAO,EAAE,gBAAgB,EAAE,MAAM,EAAE,aAAa,GAAG,gBAAgB,CAoChG"}

package/dist/distiller/payload-privacy.js

@@ -0,0 +1,51 @@
+/**
+ * Second-pass privacy over the structured distilled payload (spec §4.3).
+ *
+ * The first pass (RawMessagePrivacyProcessor) redacts raw message text before
+ * the LLM sees it; this second pass runs right before the payload is written to
+ * the outbox, catching sensitive strings the model might have reproduced or
+ * synthesized in its OUTPUT (e.g. reassembling a phone number from context).
+ *
+ * Field policy mirrors PrivacyProcessor (privacy.ts): free-text narrative fields
+ * are redacted; structural fields (type, authority, status, evidence, dates,
+ * enums), entity NAMES (aligned with Entity.name never being redacted) and
+ * reference urls (the url is the core field, checked against evidence) are left
+ * intact.
+ */
+import { RawMessagePrivacyProcessor } from "./raw-privacy.js";
+export function redactPayload(payload, config) {
+    if (!config.enabled)
+        return payload;
+    // Force irreversible for the structured pass: the reversible restoration map
+    // is only kept for raw message text keyed by msg_id (first pass); structured
+    // output redactions are not reversible (nothing to anchor them to).
+    const proc = new RawMessagePrivacyProcessor({ ...config, mode: "irreversible" });
+    const r = (text) => proc.redactString(text);
+    const rOpt = (text) => (text === undefined ? undefined : r(text));
+    const redactSignal = (sig) => {
+        const common = {
+            topic: r(sig.topic),
+            what: r(sig.what),
+            why: rOpt(sig.why),
+            project: sig.project,
+            entities: sig.entities,
+            authority: sig.authority,
+            supersedes_topic: rOpt(sig.supersedes_topic),
+            evidence: sig.evidence,
+            persistence_reason: r(sig.persistence_reason),
+        };
+        switch (sig.type) {
+            case "task":
+                return { ...sig, ...common, owner: rOpt(sig.owner) };
+            case "reference":
+                return { ...sig, ...common, trigger: rOpt(sig.trigger) };
+            case "preference":
+                return { ...sig, ...common, subject: r(sig.subject) };
+            default:
+                // decision / knowledge / discovery carry no extra free-text fields.
+                return { ...sig, ...common };
+        }
+    };
+    return { signals: payload.signals.map(redactSignal) };
+}
+//# sourceMappingURL=payload-privacy.js.map

package/dist/distiller/payload-privacy.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"payload-privacy.js","sourceRoot":"","sources":["../../src/distiller/payload-privacy.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;GAaG;AAIH,OAAO,EAAE,0BAA0B,EAAE,MAAM,kBAAkB,CAAC;AAE9D,MAAM,UAAU,aAAa,CAAC,OAAyB,EAAE,MAAqB;IAC5E,IAAI,CAAC,MAAM,CAAC,OAAO;QAAE,OAAO,OAAO,CAAC;IAEpC,6EAA6E;IAC7E,6EAA6E;IAC7E,oEAAoE;IACpE,MAAM,IAAI,GAAG,IAAI,0BAA0B,CAAC,EAAE,GAAG,MAAM,EAAE,IAAI,EAAE,cAAc,EAAE,CAAC,CAAC;IACjF,MAAM,CAAC,GAAG,CAAC,IAAY,EAAE,EAAE,CAAC,IAAI,CAAC,YAAY,CAAC,IAAI,CAAC,CAAC;IACpD,MAAM,IAAI,GAAG,CAAC,IAAwB,EAAE,EAAE,CAAC,CAAC,IAAI,KAAK,SAAS,CAAC,CAAC,CAAC,SAAS,CAAC,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC;IAEtF,MAAM,YAAY,GAAG,CAAC,GAAoB,EAAmB,EAAE;QAC7D,MAAM,MAAM,GAAG;YACb,KAAK,EAAE,CAAC,CAAC,GAAG,CAAC,KAAK,CAAC;YACnB,IAAI,EAAE,CAAC,CAAC,GAAG,CAAC,IAAI,CAAC;YACjB,GAAG,EAAE,IAAI,CAAC,GAAG,CAAC,GAAG,CAAC;YAClB,OAAO,EAAE,GAAG,CAAC,OAAO;YACpB,QAAQ,EAAE,GAAG,CAAC,QAAQ;YACtB,SAAS,EAAE,GAAG,CAAC,SAAS;YACxB,gBAAgB,EAAE,IAAI,CAAC,GAAG,CAAC,gBAAgB,CAAC;YAC5C,QAAQ,EAAE,GAAG,CAAC,QAAQ;YACtB,kBAAkB,EAAE,CAAC,CAAC,GAAG,CAAC,kBAAkB,CAAC;SAC9C,CAAC;QACF,QAAQ,GAAG,CAAC,IAAI,EAAE,CAAC;YACjB,KAAK,MAAM;gBACT,OAAO,EAAE,GAAG,GAAG,EAAE,GAAG,MAAM,EAAE,KAAK,EAAE,IAAI,CAAC,GAAG,CAAC,KAAK,CAAC,EAAE,CAAC;YACvD,KAAK,WAAW;gBACd,OAAO,EAAE,GAAG,GAAG,EAAE,GAAG,MAAM,EAAE,OAAO,EAAE,IAAI,CAAC,GAAG,CAAC,OAAO,CAAC,EAAE,CAAC;YAC3D,KAAK,YAAY;gBACf,OAAO,EAAE,GAAG,GAAG,EAAE,GAAG,MAAM,EAAE,OAAO,EAAE,CAAC,CAAC,GAAG,CAAC,OAAO,CAAC,EAAE,CAAC;YACxD;gBACE,oEAAoE;gBACpE,OAAO,EAAE,GAAG,GAAG,EAAE,GAAG,MAAM,EAAE,CAAC;QACjC,CAAC;IACH,CAAC,CAAC;IAEF,OAAO,EAAE,OAAO,EAAE,OAAO,CAAC,OAAO,CAAC,GAAG,CAAC,YAAY,CAAC,EAAE,CAAC;AACxD,CAAC"}

package/dist/distiller/prompt.d.ts

@@ -0,0 +1,28 @@
+/**
+ * Distillation prompt assembly (spec §5, §5.1, §11 task 2.9).
+ *
+ * The judgment-declaration criteria live in an embedded markdown prompt
+ * (`src/extractors/prompts/distill.md`, inlined by gen:assets into PROMPTS).
+ * This module wraps that criteria into the concrete map + reduce prompts, with
+ * carry-forward and overturned instructions threaded in.
+ *
+ * Keeping prompt text out of map-reduce.ts lets the criteria be iterated
+ * (PR-1 eval loop) without touching orchestration logic.
+ */
+import type { SegmentSummary } from "./map-reduce.js";
+/** Load the embedded distill criteria; falls back to a terse built-in if absent. */
+export declare function getDistillCriteria(): string;
+export interface MapPromptInput {
+    criteria?: string;
+    segmentText: string;
+    segNo: number;
+    carryForward: string;
+}
+export declare function getMapPrompt(input: MapPromptInput): string;
+export interface ReducePromptInput {
+    criteria?: string;
+    segmentSummaries: SegmentSummary[];
+    overturnedTopics: string[];
+}
+export declare function getReducePrompt(input: ReducePromptInput): string;
+//# sourceMappingURL=prompt.d.ts.map

package/dist/distiller/prompt.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"prompt.d.ts","sourceRoot":"","sources":["../../src/distiller/prompt.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;GAUG;AAGH,OAAO,KAAK,EAAE,cAAc,EAAE,MAAM,iBAAiB,CAAC;AAEtD,oFAAoF;AACpF,wBAAgB,kBAAkB,IAAI,MAAM,CAE3C;AAQD,MAAM,WAAW,cAAc;IAC7B,QAAQ,CAAC,EAAE,MAAM,CAAC;IAClB,WAAW,EAAE,MAAM,CAAC;IACpB,KAAK,EAAE,MAAM,CAAC;IACd,YAAY,EAAE,MAAM,CAAC;CACtB;AAED,wBAAgB,YAAY,CAAC,KAAK,EAAE,cAAc,GAAG,MAAM,CAwB1D;AAED,MAAM,WAAW,iBAAiB;IAChC,QAAQ,CAAC,EAAE,MAAM,CAAC;IAClB,gBAAgB,EAAE,cAAc,EAAE,CAAC;IACnC,gBAAgB,EAAE,MAAM,EAAE,CAAC;CAC5B;AAED,wBAAgB,eAAe,CAAC,KAAK,EAAE,iBAAiB,GAAG,MAAM,CAuBhE"}

package/dist/distiller/prompt.js

@@ -0,0 +1,69 @@
+/**
+ * Distillation prompt assembly (spec §5, §5.1, §11 task 2.9).
+ *
+ * The judgment-declaration criteria live in an embedded markdown prompt
+ * (`src/extractors/prompts/distill.md`, inlined by gen:assets into PROMPTS).
+ * This module wraps that criteria into the concrete map + reduce prompts, with
+ * carry-forward and overturned instructions threaded in.
+ *
+ * Keeping prompt text out of map-reduce.ts lets the criteria be iterated
+ * (PR-1 eval loop) without touching orchestration logic.
+ */
+import { PROMPTS } from "../embedded-assets.generated.js";
+/** Load the embedded distill criteria; falls back to a terse built-in if absent. */
+export function getDistillCriteria() {
+    return PROMPTS["distill.md"] ?? FALLBACK_CRITERIA;
+}
+const FALLBACK_CRITERIA = `You are distilling an agent session for a future retriever 30 days from now.
+Record only what a colleague would need to reconstruct WHY things are the way they
+are. Distinguish: the user deciding (user_confirmed) vs. you proposing
+(assistant_proposed) vs. a claim proven by a tool result (assistant_claimed).
+Skip transient debugging chatter and unconfirmed proposals.`;
+export function getMapPrompt(input) {
+    const criteria = input.criteria ?? getDistillCriteria();
+    const carrySection = input.carryForward.trim().length > 0
+        ? `\n## Carried forward from the previous segment\n\n${input.carryForward}\n`
+        : "";
+    return `${criteria}
+
+## Segment ${input.segNo}
+${carrySection}
+## Messages
+
+${input.segmentText}
+
+## Instructions
+
+Distill THIS segment only. Output ONLY JSON of shape:
+{
+  "seg_no": ${input.segNo},
+  "summary": "<one-paragraph gist of this segment>",
+  "tentative_signals": [ /* candidate signals per the contract */ ],
+  "overturned": [ { "topic": "<a topic from earlier that is now retracted>", "reason": "..." } ],
+  "carry_forward": "<in-progress topics / undecided items to hand to the next segment>"
+}`;
+}
+export function getReducePrompt(input) {
+    const criteria = input.criteria ?? getDistillCriteria();
+    const summariesJson = JSON.stringify(input.segmentSummaries, null, 2);
+    const overturnedList = input.overturnedTopics.length > 0
+        ? input.overturnedTopics.map((t) => `- ${t}`).join("\n")
+        : "(none)";
+    return `${criteria}
+
+## Segment summaries
+
+${summariesJson}
+
+## Overturned topics (MUST NOT appear in the final signals)
+
+${overturnedList}
+
+## Instructions
+
+Merge the tentative signals across all segments into the final session payload.
+RULE: any conclusion that appears in the overturned list above MUST NOT enter the
+final signals. Deduplicate: no two final signals may share the same (type, topic).
+Output ONLY JSON of shape { "signals": [ ...signals per the contract... ] }.`;
+}
+//# sourceMappingURL=prompt.js.map

package/dist/distiller/prompt.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"prompt.js","sourceRoot":"","sources":["../../src/distiller/prompt.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;GAUG;AAEH,OAAO,EAAE,OAAO,EAAE,MAAM,iCAAiC,CAAC;AAG1D,oFAAoF;AACpF,MAAM,UAAU,kBAAkB;IAChC,OAAO,OAAO,CAAC,YAAY,CAAC,IAAI,iBAAiB,CAAC;AACpD,CAAC;AAED,MAAM,iBAAiB,GAAG;;;;4DAIkC,CAAC;AAS7D,MAAM,UAAU,YAAY,CAAC,KAAqB;IAChD,MAAM,QAAQ,GAAG,KAAK,CAAC,QAAQ,IAAI,kBAAkB,EAAE,CAAC;IACxD,MAAM,YAAY,GAChB,KAAK,CAAC,YAAY,CAAC,IAAI,EAAE,CAAC,MAAM,GAAG,CAAC;QAClC,CAAC,CAAC,qDAAqD,KAAK,CAAC,YAAY,IAAI;QAC7E,CAAC,CAAC,EAAE,CAAC;IACT,OAAO,GAAG,QAAQ;;aAEP,KAAK,CAAC,KAAK;EACtB,YAAY;;;EAGZ,KAAK,CAAC,WAAW;;;;;;cAML,KAAK,CAAC,KAAK;;;;;EAKvB,CAAC;AACH,CAAC;AAQD,MAAM,UAAU,eAAe,CAAC,KAAwB;IACtD,MAAM,QAAQ,GAAG,KAAK,CAAC,QAAQ,IAAI,kBAAkB,EAAE,CAAC;IACxD,MAAM,aAAa,GAAG,IAAI,CAAC,SAAS,CAAC,KAAK,CAAC,gBAAgB,EAAE,IAAI,EAAE,CAAC,CAAC,CAAC;IACtE,MAAM,cAAc,GAClB,KAAK,CAAC,gBAAgB,CAAC,MAAM,GAAG,CAAC;QAC/B,CAAC,CAAC,KAAK,CAAC,gBAAgB,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,KAAK,CAAC,EAAE,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC;QACxD,CAAC,CAAC,QAAQ,CAAC;IACf,OAAO,GAAG,QAAQ;;;;EAIlB,aAAa;;;;EAIb,cAAc;;;;;;;6EAO6D,CAAC;AAC9E,CAAC"}

package/dist/distiller/raw-privacy.d.ts

@@ -0,0 +1,55 @@
+/**
+ * RawMessagePrivacyProcessor — pre-LLM redaction (spec §4.3, first pass).
+ *
+ * The legacy block pipeline sends raw messages to the LLM and only redacts the
+ * structured ExtractionResult afterwards (privacy.ts / PrivacyProcessor) — so
+ * the model sees un-redacted originals. For the distiller we fix that ordering:
+ * this processor redacts raw message TEXT before it is ever handed to the LLM
+ * (every map/reduce segment input goes through it first).
+ *
+ * It reuses the exact PrivacyConfig rules and patterns as PrivacyProcessor, but
+ * operates on plain strings keyed by msg_id (not on ExtractionResult). In
+ * reversible mode the restoration map is grouped by msg_id (spec §4.3) so an
+ * authorized reader can recover the original text for a cited evidence range.
+ */
+import type { PrivacyConfig } from "../core/config.js";
+import type { ParsedMessage } from "./msg-id.js";
+/** A single reversible redaction: what was replaced, with what, and where. */
+export interface RawRedactionEntry {
+    original: string;
+    replacement: string;
+    /** Character offset in the redacted string (post-substitution). */
+    position: number;
+}
+/** Restoration map: msg_id → the redactions applied to that message. */
+export type RestorationMap = Record<string, RawRedactionEntry[]>;
+export interface RedactMessagesResult {
+    messages: ParsedMessage[];
+    restorationMap: RestorationMap;
+}
+export declare class RawMessagePrivacyProcessor {
+    private readonly config;
+    constructor(config: PrivacyConfig);
+    private patterns;
+    private escapeRegex;
+    /**
+     * Redact one message's text. Returns the redacted string and (in reversible
+     * mode) the ordered redaction entries.
+     */
+    private redactText;
+    /**
+     * Redact a single free-text string (no restoration tracking exposed).
+     * Used by the second-pass payload privacy (payload-privacy.ts).
+     */
+    redactString(text: string): string;
+    /** Redact every message's text before it is sent to the LLM. */
+    redactMessages(messages: ParsedMessage[]): RedactMessagesResult;
+    /**
+     * Reverse redaction for a single message's (redacted) text using its
+     * msg_id-keyed entries. Applies entries in forward order (they were recorded
+     * against post-substitution offsets in reverse, so left-to-right restore is
+     * position-stable because each restore preserves earlier offsets).
+     */
+    restore(msgId: string, redactedText: string, map: RestorationMap): string;
+}
+//# sourceMappingURL=raw-privacy.d.ts.map

package/dist/distiller/raw-privacy.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"raw-privacy.d.ts","sourceRoot":"","sources":["../../src/distiller/raw-privacy.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;GAaG;AAEH,OAAO,KAAK,EAAE,aAAa,EAAE,MAAM,mBAAmB,CAAC;AACvD,OAAO,KAAK,EAAE,aAAa,EAAE,MAAM,aAAa,CAAC;AAEjD,8EAA8E;AAC9E,MAAM,WAAW,iBAAiB;IAChC,QAAQ,EAAE,MAAM,CAAC;IACjB,WAAW,EAAE,MAAM,CAAC;IACpB,mEAAmE;IACnE,QAAQ,EAAE,MAAM,CAAC;CAClB;AAED,wEAAwE;AACxE,MAAM,MAAM,cAAc,GAAG,MAAM,CAAC,MAAM,EAAE,iBAAiB,EAAE,CAAC,CAAC;AAEjE,MAAM,WAAW,oBAAoB;IACnC,QAAQ,EAAE,aAAa,EAAE,CAAC;IAC1B,cAAc,EAAE,cAAc,CAAC;CAChC;AAQD,qBAAa,0BAA0B;IACzB,OAAO,CAAC,QAAQ,CAAC,MAAM;gBAAN,MAAM,EAAE,aAAa;IAGlD,OAAO,CAAC,QAAQ;IAchB,OAAO,CAAC,WAAW;IAInB;;;OAGG;IACH,OAAO,CAAC,UAAU;IAsClB;;;OAGG;IACH,YAAY,CAAC,IAAI,EAAE,MAAM,GAAG,MAAM;IAKlC,gEAAgE;IAChE,cAAc,CAAC,QAAQ,EAAE,aAAa,EAAE,GAAG,oBAAoB;IAe/D;;;;;OAKG;IACH,OAAO,CAAC,KAAK,EAAE,MAAM,EAAE,YAAY,EAAE,MAAM,EAAE,GAAG,EAAE,cAAc,GAAG,MAAM;CAc1E"}

package/dist/distiller/raw-privacy.js

@@ -0,0 +1,123 @@
+/**
+ * RawMessagePrivacyProcessor — pre-LLM redaction (spec §4.3, first pass).
+ *
+ * The legacy block pipeline sends raw messages to the LLM and only redacts the
+ * structured ExtractionResult afterwards (privacy.ts / PrivacyProcessor) — so
+ * the model sees un-redacted originals. For the distiller we fix that ordering:
+ * this processor redacts raw message TEXT before it is ever handed to the LLM
+ * (every map/reduce segment input goes through it first).
+ *
+ * It reuses the exact PrivacyConfig rules and patterns as PrivacyProcessor, but
+ * operates on plain strings keyed by msg_id (not on ExtractionResult). In
+ * reversible mode the restoration map is grouped by msg_id (spec §4.3) so an
+ * authorized reader can recover the original text for a cited evidence range.
+ */
+export class RawMessagePrivacyProcessor {
+    config;
+    constructor(config) {
+        this.config = config;
+    }
+    // Patterns mirror PrivacyProcessor (privacy.ts) — same PrivacyConfig rules.
+    patterns() {
+        return [
+            { regex: /1[3-9]\d{9}/g, replacement: "[REDACTED_PHONE]", enabled: this.config.redact_phone },
+            { regex: /\d{17}[\dXx]/g, replacement: "[REDACTED_ID]", enabled: this.config.redact_id_card },
+            {
+                regex: /\d{16,19}/g,
+                replacement: "[REDACTED_CARD]",
+                enabled: this.config.redact_bank_card,
+            },
+            // L2 IP address — always on in PrivacyProcessor.
+            { regex: /\d{1,3}\.\d{1,3}\.\d{1,3}\.\d{1,3}/g, replacement: "[REDACTED_IP]", enabled: true },
+        ];
+    }
+    escapeRegex(str) {
+        return str.replace(/[.*+?^${}()|[\]\\]/g, "\\$&");
+    }
+    /**
+     * Redact one message's text. Returns the redacted string and (in reversible
+     * mode) the ordered redaction entries.
+     */
+    redactText(text) {
+        const entries = [];
+        let result = text;
+        const applyPattern = (regex, replacement) => {
+            const matches = [];
+            const source = result;
+            let m;
+            // biome-ignore lint/suspicious/noAssignInExpressions: standard regex.exec loop
+            while ((m = regex.exec(source)) !== null) {
+                matches.push({ original: m[0], position: m.index });
+            }
+            regex.lastIndex = 0;
+            // Apply in reverse to keep positions valid.
+            for (let i = matches.length - 1; i >= 0; i--) {
+                const mm = matches[i];
+                result =
+                    result.slice(0, mm.position) +
+                        replacement +
+                        result.slice(mm.position + mm.original.length);
+                if (this.config.mode === "reversible") {
+                    entries.push({ original: mm.original, replacement, position: mm.position });
+                }
+            }
+        };
+        for (const p of this.patterns()) {
+            if (p.enabled)
+                applyPattern(p.regex, p.replacement);
+        }
+        // L3 blocked words (case-insensitive).
+        for (const word of this.config.blocked_words) {
+            const regex = new RegExp(`\\b${this.escapeRegex(word)}\\b`, "gi");
+            applyPattern(regex, this.config.replacement);
+        }
+        return { text: result, entries };
+    }
+    /**
+     * Redact a single free-text string (no restoration tracking exposed).
+     * Used by the second-pass payload privacy (payload-privacy.ts).
+     */
+    redactString(text) {
+        if (!this.config.enabled)
+            return text;
+        return this.redactText(text).text;
+    }
+    /** Redact every message's text before it is sent to the LLM. */
+    redactMessages(messages) {
+        if (!this.config.enabled) {
+            return { messages, restorationMap: {} };
+        }
+        const restorationMap = {};
+        const redacted = messages.map((msg) => {
+            const { text, entries } = this.redactText(msg.content);
+            if (this.config.mode === "reversible" && entries.length > 0) {
+                restorationMap[msg.msgId] = entries;
+            }
+            return { ...msg, content: text };
+        });
+        return { messages: redacted, restorationMap };
+    }
+    /**
+     * Reverse redaction for a single message's (redacted) text using its
+     * msg_id-keyed entries. Applies entries in forward order (they were recorded
+     * against post-substitution offsets in reverse, so left-to-right restore is
+     * position-stable because each restore preserves earlier offsets).
+     */
+    restore(msgId, redactedText, map) {
+        const entries = map[msgId];
+        if (!entries || entries.length === 0)
+            return redactedText;
+        // Restore from rightmost to leftmost so earlier offsets stay valid.
+        const sorted = [...entries].sort((a, b) => b.position - a.position);
+        let result = redactedText;
+        for (const e of sorted) {
+            const at = result.indexOf(e.replacement, e.position);
+            const idx = at >= 0 ? at : result.indexOf(e.replacement);
+            if (idx < 0)
+                continue;
+            result = result.slice(0, idx) + e.original + result.slice(idx + e.replacement.length);
+        }
+        return result;
+    }
+}
+//# sourceMappingURL=raw-privacy.js.map

package/dist/distiller/raw-privacy.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"raw-privacy.js","sourceRoot":"","sources":["../../src/distiller/raw-privacy.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;GAaG;AA2BH,MAAM,OAAO,0BAA0B;IACR;IAA7B,YAA6B,MAAqB;QAArB,WAAM,GAAN,MAAM,CAAe;IAAG,CAAC;IAEtD,4EAA4E;IACpE,QAAQ;QACd,OAAO;YACL,EAAE,KAAK,EAAE,cAAc,EAAE,WAAW,EAAE,kBAAkB,EAAE,OAAO,EAAE,IAAI,CAAC,MAAM,CAAC,YAAY,EAAE;YAC7F,EAAE,KAAK,EAAE,eAAe,EAAE,WAAW,EAAE,eAAe,EAAE,OAAO,EAAE,IAAI,CAAC,MAAM,CAAC,cAAc,EAAE;YAC7F;gBACE,KAAK,EAAE,YAAY;gBACnB,WAAW,EAAE,iBAAiB;gBAC9B,OAAO,EAAE,IAAI,CAAC,MAAM,CAAC,gBAAgB;aACtC;YACD,iDAAiD;YACjD,EAAE,KAAK,EAAE,qCAAqC,EAAE,WAAW,EAAE,eAAe,EAAE,OAAO,EAAE,IAAI,EAAE;SAC9F,CAAC;IACJ,CAAC;IAEO,WAAW,CAAC,GAAW;QAC7B,OAAO,GAAG,CAAC,OAAO,CAAC,qBAAqB,EAAE,MAAM,CAAC,CAAC;IACpD,CAAC;IAED;;;OAGG;IACK,UAAU,CAAC,IAAY;QAC7B,MAAM,OAAO,GAAwB,EAAE,CAAC;QACxC,IAAI,MAAM,GAAG,IAAI,CAAC;QAElB,MAAM,YAAY,GAAG,CAAC,KAAa,EAAE,WAAmB,EAAE,EAAE;YAC1D,MAAM,OAAO,GAAkD,EAAE,CAAC;YAClE,MAAM,MAAM,GAAG,MAAM,CAAC;YACtB,IAAI,CAAyB,CAAC;YAC9B,+EAA+E;YAC/E,OAAO,CAAC,CAAC,GAAG,KAAK,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC,KAAK,IAAI,EAAE,CAAC;gBACzC,OAAO,CAAC,IAAI,CAAC,EAAE,QAAQ,EAAE,CAAC,CAAC,CAAC,CAAC,EAAE,QAAQ,EAAE,CAAC,CAAC,KAAK,EAAE,CAAC,CAAC;YACtD,CAAC;YACD,KAAK,CAAC,SAAS,GAAG,CAAC,CAAC;YACpB,4CAA4C;YAC5C,KAAK,IAAI,CAAC,GAAG,OAAO,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC,IAAI,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC;gBAC7C,MAAM,EAAE,GAAG,OAAO,CAAC,CAAC,CAAC,CAAC;gBACtB,MAAM;oBACJ,MAAM,CAAC,KAAK,CAAC,CAAC,EAAE,EAAE,CAAC,QAAQ,CAAC;wBAC5B,WAAW;wBACX,MAAM,CAAC,KAAK,CAAC,EAAE,CAAC,QAAQ,GAAG,EAAE,CAAC,QAAQ,CAAC,MAAM,CAAC,CAAC;gBACjD,IAAI,IAAI,CAAC,MAAM,CAAC,IAAI,KAAK,YAAY,EAAE,CAAC;oBACtC,OAAO,CAAC,IAAI,CAAC,EAAE,QAAQ,EAAE,EAAE,CAAC,QAAQ,EAAE,WAAW,EAAE,QAAQ,EAAE,EAAE,CAAC,QAAQ,EAAE,CAAC,CAAC;gBAC9E,CAAC;YACH,CAAC;QACH,CAAC,CAAC;QAEF,KAAK,MAAM,CAAC,IAAI,IAAI,CAAC,QAAQ,EAAE,EAAE,CAAC;YAChC,IAAI,CAAC,CAAC,OAAO;gBAAE,YAAY,CAAC,CAAC,CAAC,KAAK,EAAE,CAAC,CAAC,WAAW,CAAC,CAAC;QACtD,CAAC;QACD,uCAAuC;QACvC,KAAK,MAAM,IAAI,IAAI,IAAI,CAAC,MAAM,CAAC,aAAa,EAAE,CAAC;YAC7C,MAAM,KAAK,GAAG,IAAI,MAAM,CAAC,MAAM,IAAI,CAAC,WAAW,CAAC,IAAI,CAAC,KAAK,EAAE,IAAI,CAAC,CAAC;YAClE,YAAY,CAAC,KAAK,EAAE,IAAI,CAAC,MAAM,CAAC,WAAW,CAAC,CAAC;QAC/C,CAAC;QAED,OAAO,EAAE,IAAI,EAAE,MAAM,EAAE,OAAO,EAAE,CAAC;IACnC,CAAC;IAED;;;OAGG;IACH,YAAY,CAAC,IAAY;QACvB,IAAI,CAAC,IAAI,CAAC,MAAM,CAAC,OAAO;YAAE,OAAO,IAAI,CAAC;QACtC,OAAO,IAAI,CAAC,UAAU,CAAC,IAAI,CAAC,CAAC,IAAI,CAAC;IACpC,CAAC;IAED,gEAAgE;IAChE,cAAc,CAAC,QAAyB;QACtC,IAAI,CAAC,IAAI,CAAC,MAAM,CAAC,OAAO,EAAE,CAAC;YACzB,OAAO,EAAE,QAAQ,EAAE,cAAc,EAAE,EAAE,EAAE,CAAC;QAC1C,CAAC;QACD,MAAM,cAAc,GAAmB,EAAE,CAAC;QAC1C,MAAM,QAAQ,GAAG,QAAQ,CAAC,GAAG,CAAC,CAAC,GAAG,EAAE,EAAE;YACpC,MAAM,EAAE,IAAI,EAAE,OAAO,EAAE,GAAG,IAAI,CAAC,UAAU,CAAC,GAAG,CAAC,OAAO,CAAC,CAAC;YACvD,IAAI,IAAI,CAAC,MAAM,CAAC,IAAI,KAAK,YAAY,IAAI,OAAO,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;gBAC5D,cAAc,CAAC,GAAG,CAAC,KAAK,CAAC,GAAG,OAAO,CAAC;YACtC,CAAC;YACD,OAAO,EAAE,GAAG,GAAG,EAAE,OAAO,EAAE,IAAI,EAAE,CAAC;QACnC,CAAC,CAAC,CAAC;QACH,OAAO,EAAE,QAAQ,EAAE,QAAQ,EAAE,cAAc,EAAE,CAAC;IAChD,CAAC;IAED;;;;;OAKG;IACH,OAAO,CAAC,KAAa,EAAE,YAAoB,EAAE,GAAmB;QAC9D,MAAM,OAAO,GAAG,GAAG,CAAC,KAAK,CAAC,CAAC;QAC3B,IAAI,CAAC,OAAO,IAAI,OAAO,CAAC,MAAM,KAAK,CAAC;YAAE,OAAO,YAAY,CAAC;QAC1D,oEAAoE;QACpE,MAAM,MAAM,GAAG,CAAC,GAAG,OAAO,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,QAAQ,GAAG,CAAC,CAAC,QAAQ,CAAC,CAAC;QACpE,IAAI,MAAM,GAAG,YAAY,CAAC;QAC1B,KAAK,MAAM,CAAC,IAAI,MAAM,EAAE,CAAC;YACvB,MAAM,EAAE,GAAG,MAAM,CAAC,OAAO,CAAC,CAAC,CAAC,WAAW,EAAE,CAAC,CAAC,QAAQ,CAAC,CAAC;YACrD,MAAM,GAAG,GAAG,EAAE,IAAI,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,MAAM,CAAC,OAAO,CAAC,CAAC,CAAC,WAAW,CAAC,CAAC;YACzD,IAAI,GAAG,GAAG,CAAC;gBAAE,SAAS;YACtB,MAAM,GAAG,MAAM,CAAC,KAAK,CAAC,CAAC,EAAE,GAAG,CAAC,GAAG,CAAC,CAAC,QAAQ,GAAG,MAAM,CAAC,KAAK,CAAC,GAAG,GAAG,CAAC,CAAC,WAAW,CAAC,MAAM,CAAC,CAAC;QACxF,CAAC;QACD,OAAO,MAAM,CAAC;IAChB,CAAC;CACF"}