Skip to content

fix(engine): reject malformed historical replay calibration scores #3466

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
JSONbored merged 1 commit into from
Jul 5, 2026
Merged

fix(engine): reject malformed historical replay calibration scores #3466

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
12 changes: 9 additions & 3 deletions packages/gittensory-engine/src/phase7-calibration-loop.ts
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -331,9 +331,11 @@ export function shouldScheduleHistoricalReplayRun(input: {

function extractHistoricalReplayScore(
compositeScore: number | GateVerdictCompositeCalibrationScore,
): number {
if (typeof compositeScore === "number") return roundScore(compositeScore);
return roundScore(compositeScore.compositeScore);
): number | null {
const rawScore =
typeof compositeScore === "number" ? compositeScore : compositeScore.compositeScore;
if (!Number.isFinite(rawScore)) return null;
return roundScore(rawScore);
}

/**
Expand Down Expand Up @@ -423,6 +425,10 @@ export function computePhase7CalibrationLoop(input: {
replayHarnessHold = true;
holdReasons.push("replay_run_stale");
rejectedSources.push({ source: "historical_replay", reason: "replay_run_stale" });
} else if (accuracy === null) {
replayHarnessHold = true;
holdReasons.push("invalid_replay_score");
rejectedSources.push({ source: "historical_replay", reason: "invalid_replay_score" });
} else {
contributingSources.push("historical_replay");
}
Expand Down
22 changes: 22 additions & 0 deletions packages/gittensory-engine/test/phase7-calibration-loop.test.ts
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -345,6 +345,28 @@ test("computePhase7CalibrationLoop fails closed when the replay harness is degra
}
});

test("REGRESSION: malformed historical replay scores fail closed instead of permitting autonomy increases", () => {
for (const compositeScore of [{} as never, Number.NaN, Number.POSITIVE_INFINITY] as const) {
const result = computePhase7CalibrationLoop({
config: enabledConfig({ autonomyIncreaseMinAccuracy: 0.99 }),
prOutcome: sufficientPrOutcome(1),
historicalReplay: {
...healthyReplay(0.82),
compositeScore,
},
now: NOW,
});

assert.equal(result.bySource.historical_replay.accuracy, null);
assert.equal(result.combinedAccuracy, 1);
assert.equal(result.replayHarnessHold, true);
assert.equal(result.autonomyIncreasePermitted, false);
assert.ok(result.holdReasons.includes("invalid_replay_score"));
assert.ok(!result.audit.contributingSources.includes("historical_replay"));
assert.ok(result.audit.rejectedSources.some((row) => row.reason === "invalid_replay_score"));
}
});

test("computePhase7CalibrationLoop fails closed on stale replay rather than silently using pr_outcome only", () => {
const result = computePhase7CalibrationLoop({
config: enabledConfig(),
Expand Down