fix(rectification): wire unique-minute confirmation to VedAstro and sealed holdout

Keep confirmation fail-closed until the public AA set is ready, and rewrite Technique Audit from the attached VedAstro status instead of a hardcoded blocked row.

Co-authored-by: Cursor <cursoragent@cursor.com>
This commit is contained in:
Jesse_Chen
2026-08-20 10:05:35 +08:00
co-authored by Cursor
parent a93a3cc1a9
commit 55d3119c68
34 changed files with 1220 additions and 59 deletions
@@ -201,5 +201,7 @@ test("adopted minute recasts the house table and keeps technique audit collapsed
assert.equal(result?.techniqueAudit.length, 3);
assert.equal(result?.techniqueAudit[2]?.status, "blocked");
assert.equal(result?.precisionStage?.current, "d4_refine");
assert.equal(result?.confirmationGate.confirmation_allowed, false);
assert.equal(result?.confirmationGate.blockers.some((item) => item.id === "public_aa_holdout" && item.status === "not_ready"), true);
});
@@ -71,10 +71,16 @@ function blocker(
}
test("sealed holdout aggregates match the v2 report and stay below the case gate", () => {
const productHoldout = JSON.parse(readFileSync(
new URL("../../references/rectification_sealed_holdout.v1.json", import.meta.url),
"utf8",
)) as typeof SEALED_MINUTE_HOLDOUT;
assert.equal(SEALED_MINUTE_HOLDOUT.valid_public_aa_cases, holdoutReport.case_gate.valid_public_aa_cases);
assert.equal(SEALED_MINUTE_HOLDOUT.required_cases, holdoutReport.case_gate.minimum_public_aa_cases);
assert.equal(SEALED_MINUTE_HOLDOUT.top_1_rate, holdoutReport.metrics.top_1_rate);
assert.equal(SEALED_MINUTE_HOLDOUT.confirmation_coverage_rate, holdoutReport.metrics.confirmation_coverage_rate);
assert.equal(SEALED_MINUTE_HOLDOUT.valid_public_aa_cases, productHoldout.valid_public_aa_cases);
assert.equal(SEALED_MINUTE_HOLDOUT.status, productHoldout.status);
assert.ok(SEALED_MINUTE_HOLDOUT.valid_public_aa_cases < SEALED_MINUTE_HOLDOUT.required_cases);
assert.equal(SEALED_MINUTE_HOLDOUT.status, "not_ready");
});
@@ -158,7 +164,7 @@ test("holdout not_ready forbids unique-minute copy and still blocks confirm", as
assert.match(agentSource, /session_outcome=adopt_representative/);
assert.doesNotMatch(agentSource, /±2 分钟/);
assert.equal(PUBLIC_RECTIFICATION_TOOLS.length, 13);
assert.equal(RECTIFICATION_SKILL_VERSION, "10.0.5");
assert.equal(RECTIFICATION_SKILL_VERSION, "10.0.6");
const accounting = fakeAccounting({
...receiptHandlers,
@@ -504,9 +504,9 @@ test("rescore failure does not fail the evidence write", async () => {
assert.ok(result.rescore.error_code);
});
test("public tool surface stays at 13 and new cases bind 10.0.5", () => {
test("public tool surface stays at 13 and new cases bind 10.0.6", () => {
assert.equal(PUBLIC_RECTIFICATION_TOOLS.length, 13);
assert.equal(RECTIFICATION_SKILL_VERSION, "10.0.5");
assert.equal(RECTIFICATION_SKILL_VERSION, "10.0.6");
const deprecated = resolveExactSkillPackage(
"jyotish-birth-time-rectification",
"10.0.2",
@@ -202,9 +202,9 @@ test("read-case evidence context keeps day labels and confirm does not rewrite d
assert.equal("p_occurred_from" in confirmCall.args, false);
});
test("new-case skill identity is 10.0.5 and the prompt prefers batch ingest", () => {
assert.equal(RECTIFICATION_SKILL_VERSION, "10.0.5");
assert.match(skill, /^version: 10\.0\.5$/m);
test("new-case skill identity is 10.0.6 and the prompt prefers batch ingest", () => {
assert.equal(RECTIFICATION_SKILL_VERSION, "10.0.6");
assert.match(skill, /^version: 10\.0\.6$/m);
assert.match(skill, /不要对同一句用户消息里的多件事件逐条 propose\+confirm/);
assert.match(agentSource, /当前轮新事件一律走 rectification-record-evidence-batch/);
assert.doesNotMatch(agentSource, /分别调用 rectification-propose-evidence 和 rectification-confirm-evidence/);
@@ -76,11 +76,11 @@ test("system prompt carries only high-priority boundaries, never the method copy
test("agent pins the dedicated rectification skill and its fixed version", () => {
assert.equal(RECTIFICATION_V9_SKILL_NAME, "jyotish-birth-time-rectification");
assert.equal(basename(RECTIFICATION_V9_SKILL_PATH), RECTIFICATION_V9_SKILL_NAME);
assert.ok(RECTIFICATION_V9_PACKAGE_PATH.endsWith("skills/jyotish-birth-time-rectification/versions/10.0.5"));
assert.ok(RECTIFICATION_V9_PACKAGE_PATH.endsWith("skills/jyotish-birth-time-rectification/versions/10.0.6"));
assert.notEqual(RECTIFICATION_V9_SKILL_PATH, RECTIFICATION_V9_PACKAGE_PATH);
assert.equal(realpathSync(RECTIFICATION_V9_SKILL_PATH), RECTIFICATION_V9_PACKAGE_PATH);
assert.equal(RECTIFICATION_SKILL_NAME, "jyotish-birth-time-rectification");
assert.equal(RECTIFICATION_SKILL_VERSION, "10.0.5");
assert.equal(RECTIFICATION_SKILL_VERSION, "10.0.6");
});
test("step budgets are bounded per action with a hard ceiling", () => {
@@ -92,9 +92,9 @@ test("terminal transitions are one-way and evidence writes stop at terminal", ()
test("the active rectification skill pins the v10 identity and lives in the right directory", () => {
assert.equal(RECTIFICATION_SKILL_NAME, "jyotish-birth-time-rectification");
assert.equal(RECTIFICATION_SKILL_VERSION, "10.0.5");
assert.equal(RECTIFICATION_SKILL_VERSION, "10.0.6");
assert.match(skill, /^---\nname: jyotish-birth-time-rectification/m);
assert.match(skill, /^version: 10\.0\.5$/m);
assert.match(skill, /^version: 10\.0\.6$/m);
for (const reference of references) {
const content = readFileSync(`${skillDirectory}/references/${reference}`, "utf8");
assert.ok(content.length > 0, `${reference} must be non-empty`);
@@ -3,8 +3,10 @@ import test from "node:test";
import {
RectificationEngineError,
mergeVedastroValidateIntoReceipt,
runV9CandidateScore,
runV9Diagnostics,
runV9VedastroValidate,
toEngineEvents,
type V9EngineScoreResult,
} from "../src/lib/rectification-agentic/v9/engine-client.ts";
@@ -315,3 +317,63 @@ test("engine http failures surface as safe engine errors, never raw stack traces
restore();
}
});
test("vedastro-validate maps pass/fail and never grants unique-minute confirmation", async () => {
const restore = stubEngine({
success: true,
endpoint: "rectification_v5_vedastro_validate",
status: "pass",
passed: true,
can_confirm_exact_minute: false,
minute_sensitive_validation: { status: "pass", discriminated: true },
event_validation: {
status: "pass",
search_events_primary_supports_local_winner: true,
},
});
try {
const result = await runV9VedastroValidate({
baselineBirthSnapshot: SNAPSHOT,
candidateRange: RANGE,
events: toEngineEvents(EVIDENCE),
candidateTimes: ["04:50", "04:51"],
});
assert.equal(result.status, "passed");
assert.equal(result.canConfirmExactMinute, false);
assert.equal(result.searchEventsSupportsLocalWinner, true);
const merged = mergeVedastroValidateIntoReceipt(
{ ...DECISION_RECEIPT, gates: { exact_confirmation: { external_validation_status: "passed" } } },
result,
);
const exact = (merged.gates as { exact_confirmation: Record<string, unknown> }).exact_confirmation;
assert.equal(exact.external_validation_status, "passed");
assert.equal(
(exact.vedastro_event_validation as { search_events_primary_supports_local_winner: boolean })
.search_events_primary_supports_local_winner,
true,
);
assert.equal(merged.confirmation_allowed, false);
} finally {
restore();
}
});
test("vedastro-validate timeout stays not_evaluated and is not described as fail", async () => {
const previous = globalThis.fetch;
globalThis.fetch = (async () => {
throw new Error("timeout");
}) as unknown as typeof fetch;
try {
const result = await runV9VedastroValidate({
baselineBirthSnapshot: SNAPSHOT,
candidateRange: RANGE,
events: toEngineEvents(EVIDENCE),
candidateTimes: ["04:50", "04:51"],
});
assert.equal(result.status, "not_evaluated");
assert.notEqual(result.status, "failed");
assert.equal(result.canConfirmExactMinute, false);
} finally {
globalThis.fetch = previous;
}
});
@@ -209,7 +209,7 @@ test("open RPC passes the pinned skill and server-derived baseline only", async
session_id: SESSION_ID,
status: "draft",
should_start_opening: true,
skill_version: "10.0.5",
skill_version: "10.0.6",
};
}
return null;
@@ -247,11 +247,11 @@ test("open RPC passes the pinned skill and server-derived baseline only", async
});
assert.equal(response.disposition, "created");
assert.equal(response.shouldStartOpening, true);
assert.equal(response.skillVersion, "10.0.5");
assert.equal(response.skillVersion, "10.0.6");
const openCall = accounting.calls.find((call) => call.fn === "open_agentic_rectification_case_v2");
assert.ok(openCall);
assert.equal(openCall.args.p_skill_name, "jyotish-birth-time-rectification");
assert.equal(openCall.args.p_skill_version, "10.0.5");
assert.equal(openCall.args.p_skill_version, "10.0.6");
assert.equal(openCall.args.p_user_id, "user-1");
// The server derives the baseline; the request never carries it from the browser.
assert.equal("birth_date" in openCall.args, false);
+2 -2
View File
@@ -85,8 +85,8 @@ test("checked-in registry verifies hashed product packages and leaves consult on
[
{
name: "jyotish-birth-time-rectification",
version: "10.0.5",
sha256: "47cd353b5f7a08695007ce52d77560584300a1097ed97248f3a81372eeaca9c6",
version: "10.0.6",
sha256: "25cd80f9129217723e11a88f247ea081b2d6c9ca95c71d6d5f0633b4de6f3e8b",
},
{
name: "jyotish-personal-report",