loopforge 3.0.0 → 3.2.1

package/dist/enforcement-gate.js

@@ -0,0 +1,1071 @@
+/** Enforcement Gate — Layer 2 round-boundary runtime enforcement (v1.13).
+ *
+ * Pure-function module. Receives the verification gate's findings plus
+ * round-level context and decides whether to accept the round, reject it
+ * (force the agent to redo), or terminate the loop.
+ *
+ * This is the "runtime" that prompt-only constraint systems lack.
+ * Verification gate detects WHAT is wrong; enforcement gate decides
+ * what to DO about it.
+ *
+ * Decision semantics:
+ * - accept:    round passes all checks; advance to next round as normal.
+ * - reject:    agent's self-evaluation or round output is invalid; the
+ *              agent receives a rejection prompt and must redo the SAME
+ *              round. Round counter does NOT increment.
+ * - terminate: loop has reached an unrecoverable state; stop immediately
+ *              with stopReason "enforcement_terminated".
+ */
+import { makeEnforcementResult } from "./protocol.js";
+import { entryRound, machineProgressSeries, CHECK_SUCCESS_WITH_REMAINING_CRITERIA, CHECK_RECURRING_VIOLATION, CHECK_SUCCESS_WITHOUT_VERIFIED_EVIDENCE, CHECK_BACKTRACK_WORKSPACE_NOT_RESTORED, CHECK_REQUIRED_COMMAND_FAILED, CHECK_COMMAND_EVIDENCE_MISMATCH, CHECK_OUTCOME_SUCCESS_CONTRADICTION } from "./verification-gate.js";
+import { effectiveSuccess } from "./self-eval.js";
+import { deriveConstraintId, deriveCriterionId, deriveSubGoalId } from "./loop-compiler.js";
+import { getPolicy } from "./policy.js";
+import { STABLE_ID_RE, isRecord } from "./token-utils.js";
+import { isProcessResult } from "./round-transaction.js";
+// ═══════════════════════════════════════════════════════════════════════════
+// Helpers
+// ═══════════════════════════════════════════════════════════════════════════
+/** Extract progress_estimate from a vault entry's execution_evidence.
+ *  Returns null if no execution evidence is available. */
+function entryProgress(entry) {
+    const evidence = entry.execution_evidence;
+    if (!evidence)
+        return null;
+    const pe = evidence.progress_estimate;
+    return typeof pe === "number" ? pe : null;
+}
+/** Collect progress estimates from vault entries into a round→estimate map.
+ *  Shared by R4 (progress_stall) and R5 (progress_stall_terminal). */
+function collectProgressByRound(vaultEntries, currentRound) {
+    const progressByRound = new Map();
+    for (const entry of vaultEntries) {
+        const rnd = entryRound(entry);
+        if (rnd < 1 || rnd >= currentRound)
+            continue;
+        const pe = entryProgress(entry);
+        if (pe !== null)
+            progressByRound.set(rnd, pe);
+    }
+    return progressByRound;
+}
+/** v3.2.1: R4/R5 progress window — after a backtrack committed for this
+ *  round, the agent's redo submission is the freshest data point. The
+ *  pre-rollback history (which triggered the stall) stays in the vault and
+ *  the redo does not commit while it is being evaluated, so without this
+ *  the window never changes: a genuinely fixed redo is rejected forever and
+ *  the loop dies in a spurious terminate. A redo that is STILL stalled
+ *  keeps the window flat and hits the normal reject → terminate ladder. */
+function progressWindow(vaultEntries, currentRound, selfEval) {
+    const progressByRound = collectProgressByRound(vaultEntries, currentRound);
+    if (hasCommittedBacktrack(vaultEntries, currentRound)) {
+        const pe = selfEval.execution_evidence?.progress_estimate;
+        if (typeof pe === "number")
+            progressByRound.set(currentRound, pe);
+    }
+    return progressByRound;
+}
+// ═══════════════════════════════════════════════════════════════════════════
+// v2.10: Backtrack helpers
+// ═══════════════════════════════════════════════════════════════════════════
+/** Extract the committed round decision's result from a vault entry.
+ *  The round commit entry (`loop:<id>:r<N>:feedback`) carries the full
+ *  decision under loop_lineage.round_transaction; lineage entries carry
+ *  nothing at the top level. */
+function committedRoundResult(entry) {
+    const lineage = isRecord(entry.loop_lineage) ? entry.loop_lineage : null;
+    const rt = lineage && isRecord(lineage.round_transaction)
+        ? lineage.round_transaction
+        : null;
+    if (!rt)
+        return null;
+    return isProcessResult(rt.result) ? rt.result : null;
+}
+/** True when a backtrack already committed at or after the current round —
+ *  i.e. the loop is re-walking territory the system already rolled back
+ *  once. Guards against the reject↔backtrack deadlock on persistent stalls. */
+function hasCommittedBacktrack(vaultEntries, currentRound) {
+    return vaultEntries.some((entry) => {
+        if (entryRound(entry) < currentRound)
+            return false;
+        return committedRoundResult(entry)?.action === "backtrack";
+    });
+}
+/** Check whether a vault entry represents a "clean" round that can serve
+ *  as a safe restore point. A clean round was accepted (not rejected)
+ *  and had no error-level verification flags.
+ *
+ *  v3.2.1: judged by the COMMITTED decision and gate flags, not by the
+ *  agent's self-reported success. A progress-stall round commits with
+ *  success=false but is accepted and carries no error flags — it is a
+ *  valid backtrack target (the most common backtrack trigger is exactly a
+ *  run of such rounds). Rejected/terminated rounds never commit, and
+ *  rolled-back rounds are excluded below. */
+function isCleanRound(entry) {
+    const committed = committedRoundResult(entry);
+    // A rolled-back round (already backtracked) is not a valid restore point.
+    if (committed?.action === "backtrack")
+        return false;
+    // Rejected/terminated rounds are not clean. (They never commit in the
+    // real vault — this guards legacy or mixed-shape data.)
+    if (committed && (committed.action === "reject" || committed.action === "terminate")) {
+        return false;
+    }
+    // Rounds with error-level verification flags are not clean
+    const flags = Array.isArray(entry.verification_flags)
+        ? entry.verification_flags
+        : committed?.verificationFlags ?? [];
+    for (const f of flags) {
+        if (f !== null && typeof f === "object" &&
+            f.severity === "error")
+            return false;
+    }
+    return true;
+}
+/** Scan backwards from currentRound to find the most recent clean round.
+ *  Collects discovered_constraints from skipped rounds along the way.
+ *  Returns null when no clean round is found within maxDepth. */
+export function findSafeRestorePoint(currentRound, vaultEntries, maxDepth) {
+    // v3.2.1: exclude compile-time lineage entries — every committed round
+    // writes one (engine.ts hardcodes `success: true` on it, and it carries no
+    // round_transaction), and it precedes the feedback entry in the flat view.
+    // Treating it as a decision made every round look clean, so the restore
+    // point was always currentRound-1: dirty/rejected rounds were never
+    // skipped, depth > 1 never collected skippedDiscoveries, and the restore
+    // prompt's file list (round-coordinator) was always empty.
+    const completed = vaultEntries
+        .filter((e) => {
+        const rnd = entryRound(e);
+        return rnd >= 1 && rnd < currentRound && e.task_type !== "loop_lineage";
+    })
+        .sort((a, b) => entryRound(b) - entryRound(a)); // newest first
+    const skippedDiscoveries = [];
+    for (let depth = 1; depth <= maxDepth; depth++) {
+        const targetRound = currentRound - depth;
+        if (targetRound < 1)
+            break;
+        const entry = completed.find((e) => entryRound(e) === targetRound);
+        if (!entry)
+            continue;
+        // Collect discoveries from rounds we're about to skip (depth > 1)
+        if (depth > 1) {
+            for (let d = 1; d < depth; d++) {
+                const skippedRound = currentRound - d;
+                const skippedEntry = completed.find((e) => entryRound(e) === skippedRound);
+                if (skippedEntry) {
+                    const discovered = Array.isArray(skippedEntry.discovered_constraints)
+                        ? skippedEntry.discovered_constraints.filter((v) => typeof v === "string")
+                        : [];
+                    for (const item of discovered) {
+                        if (!skippedDiscoveries.includes(item))
+                            skippedDiscoveries.push(item);
+                    }
+                }
+            }
+        }
+        if (isCleanRound(entry)) {
+            return { round: targetRound, skippedDiscoveries };
+        }
+    }
+    return null;
+}
+/** v2.12: Git HEAD commit hash at the restore point. Read from the restore
+ *  round's committed transaction snapshot (after evidence preferred, before
+ *  evidence as fallback). Returns null when the round has no git snapshot
+ *  or no committed feedback entry. */
+export function findBacktrackTargetGitHead(restoreRound, vaultEntries) {
+    const feedback = vaultEntries.find((entry) => String(entry.task_id ?? "") === `loop:${entry.loop_id}:r${restoreRound}:feedback`);
+    const lineage = feedback && isRecord(feedback.loop_lineage)
+        ? feedback.loop_lineage
+        : null;
+    const transaction = lineage && isRecord(lineage.round_transaction)
+        ? lineage.round_transaction
+        : null;
+    const snapshot = transaction && isRecord(transaction.snapshot)
+        ? transaction.snapshot
+        : null;
+    if (!snapshot)
+        return null;
+    const evidence = Array.isArray(snapshot.afterEvidence) && snapshot.afterEvidence.length > 0
+        ? snapshot.afterEvidence
+        : Array.isArray(snapshot.beforeEvidence)
+            ? snapshot.beforeEvidence
+            : [];
+    for (const item of evidence) {
+        if (!isRecord(item) || item.provider !== "git" || !isRecord(item.data))
+            continue;
+        const head = item.data.head;
+        if (typeof head === "string" && head.length > 0)
+            return head;
+    }
+    return null;
+}
+/** Build the backtrack prompt injected at the top of the restored round.
+ *
+ *  v2.13: Includes concrete workspace restore instructions with affected
+ *  file lists from skipped rounds. The agent must restore the working tree
+ *  to the clean round's state before proceeding. */
+export function buildBacktrackPrompt(fromRound, toRound, triggerRule, skippedDiscoveries, 
+/** v2.13: Files changed in the skipped rounds (from evidence snapshots).
+ *  Used to show the agent exactly what needs to be reverted. */
+skippedFiles = [], 
+/** v2.12: Git HEAD commit hash at the backtrack point (current HEAD).
+ *  The agent must discard work back to the clean round's state. */
+gitHead) {
+    const lines = [
+        `## ⛔ Backtrack — Round ${fromRound} → Restored to Round ${toRound}`,
+        "",
+    ];
+    // ── Why This Happened ────────────────────────────────────────────────
+    if (triggerRule === "progress_stall") {
+        lines.push("### Why This Happened", "", `Progress stalled over rounds ${toRound + 1}–${fromRound}. ` +
+            `The work done in those rounds produced no verifiable progress — ` +
+            `no new files changed, no criteria met.`, "", "The approach used in those rounds **did not work**. It should not be repeated.", "");
+    }
+    else if (triggerRule === "progress_stall_terminal") {
+        lines.push("### Why This Happened", "", `Progress was completely flat for multiple rounds leading up to Round ${fromRound}. ` +
+            `The agent made **zero forward motion** — the task is not advancing.`, "", "A **radically different** strategy is needed. The previous approach produced nothing.", "");
+    }
+    // ── v2.13: Workspace Restore ──────────────────────────────────────────
+    lines.push("### ⚠️ Workspace Restore Required", "", `Your working directory still contains changes from the **failed** ` +
+        `rounds (${toRound + 1}–${fromRound}). You MUST discard those changes ` +
+        `before starting Round ${toRound + 1}. Working on top of stale changes ` +
+        `will cause the verification gate to reject your next submission.`, "");
+    if (skippedFiles.length > 0) {
+        lines.push("**Files modified in skipped rounds (must be reverted):**", "");
+        for (const f of skippedFiles.slice(0, 15)) {
+            lines.push(`- \`${f}\``);
+        }
+        if (skippedFiles.length > 15) {
+            lines.push(`- … and ${skippedFiles.length - 15} more files`);
+        }
+        lines.push("");
+    }
+    lines.push("**To restore a clean workspace, run:**", "", "```bash", "# Option A: Discard all uncommitted changes (recommended)", "git checkout -- .", "git clean -fd", "", "# Option B: If you have unrelated work to keep", `git stash push -m "backtrack-safety-net-round-${fromRound}"`, "git checkout -- .", "git clean -fd", "```", "");
+    // v2.12: Restore to the exact commit observed at the clean round.
+    if (gitHead) {
+        lines.push(`**Restore to commit \`${gitHead.slice(0, 12)}\`** — the verification ` +
+            "gate checks that your working tree returns to this commit before " +
+            "accepting your next submission.", "", "```bash", "# Option C: If the failed rounds created commits, reset HEAD to the restore point", `git reset --hard ${gitHead.slice(0, 12)}`, "```", "", "> Option C discards the failed rounds' commits — use it only when those", "> commits belong to the failed work.", "");
+    }
+    lines.push(`**If git is unavailable**, manually revert the files listed above ` +
+        `to their state at Round ${toRound}. Re-read the state file below — ` +
+        `it reflects the correct state at Round ${toRound}.`, "", "> 🛡️ **Verification:** The next round's evidence check will verify that " +
+        "> your workspace no longer contains the failed changes. Submissions " +
+        "> with unrestored files will be **rejected**.", "");
+    // ── What Must Change ──────────────────────────────────────────────────
+    lines.push("### What Must Change", "", "- Do **NOT** repeat the approach used in the skipped rounds.", "- Try a **different** task decomposition or technique.", "- If the current sub-goal is stuck, consider canceling it " +
+        "(`canceled_subtasks`) and working on a different one.", "- Use the `prompt_requests.confusion_points` field in your self-evaluation " +
+        "to flag anything you don't understand about the task or the blockage.", "");
+    // ── Preserved Discoveries ─────────────────────────────────────────────
+    if (skippedDiscoveries.length > 0) {
+        lines.push("### Preserved Discoveries", "", "These were discovered in the skipped rounds and remain valid:", "");
+        for (const d of skippedDiscoveries) {
+            lines.push(`- ${d}`);
+        }
+        lines.push("");
+    }
+    // ── Instructions ──────────────────────────────────────────────────────
+    lines.push("### Instructions", "", `You have been rolled back to Round ${toRound}. ` +
+        `You are now starting **Round ${toRound + 1}** with the lessons above.`, "", "1. **First**: restore your workspace using the commands above.", "2. Read the restored state below — it reflects Round " + toRound + ".", "3. The discoveries from skipped rounds (if any) have been merged " +
+        "into the active constraints.", "4. Do NOT retrace the steps that led to the stall. Choose a new path.", "");
+    return lines.join("\n");
+}
+// ═══════════════════════════════════════════════════════════════════════════
+// Individual enforcement rules — each returns EnforcementResult | null
+// Rules are ordered by priority. The first non-null result wins.
+// ═══════════════════════════════════════════════════════════════════════════
+/** R1: Agent claims success but success_criteria_remaining has items.
+ *  This is a lie — the agent must either finish the criteria or
+ *  set success=false honestly. Triggered by the verification gate's
+ *  "success_with_remaining_criteria" error flag. */
+function enforceSuccessWithRemainingCriteria(flags) {
+    const flag = flags.find((f) => f.check === CHECK_SUCCESS_WITH_REMAINING_CRITERIA && f.severity === "error");
+    if (!flag)
+        return null;
+    return makeEnforcementResult({
+        action: "reject",
+        reason: flag.detail,
+        fix_instructions: "You set success=true but success criteria remain unmet. " +
+            "Either: (a) complete the remaining criteria and re-submit your self-evaluation, " +
+            "or (b) set success=false and honestly report what remains to be done.",
+        check: "success_with_remaining_criteria",
+    });
+}
+/** R2: Same constraint violation appears in 3 consecutive rounds.
+ *  The agent is repeating the same mistake. Triggered by the verification
+ *  gate's "recurring_violation" error flag. */
+function enforceRecurringViolation(flags) {
+    const flag = flags.find((f) => f.check === CHECK_RECURRING_VIOLATION && f.severity === "error");
+    if (!flag)
+        return null;
+    return makeEnforcementResult({
+        action: "reject",
+        reason: flag.detail,
+        fix_instructions: "The same constraint violation has appeared in 3 consecutive rounds. " +
+            "You must: (a) explain WHY this violation keeps occurring, " +
+            "and (b) propose a DIFFERENT approach than what you used in the last 3 rounds. " +
+            "Do NOT retry the same strategy — it has failed 3 times.",
+        check: "recurring_violation",
+    });
+}
+/** R-EVID (v2.14): A required command failure, a test-results mismatch that
+ *  hides failures, or a self-contradictory outcome claim contradicts the
+ *  success claim before commit. The verification gate detects the
+ *  contradiction; this rule decides what to do about it — reject so the
+ *  agent redos the SAME round with honest claims. Same-check rejections
+ *  accumulate toward R6 via the per-check counter. */
+function enforceEvidenceContradiction(flags) {
+    const contradiction = flags.find((f) => f.severity === "error" &&
+        (f.check === CHECK_REQUIRED_COMMAND_FAILED ||
+            f.check === CHECK_COMMAND_EVIDENCE_MISMATCH ||
+            f.check === CHECK_OUTCOME_SUCCESS_CONTRADICTION));
+    if (!contradiction)
+        return null;
+    return makeEnforcementResult({
+        action: "reject",
+        reason: contradiction.detail,
+        fix_instructions: "A required verification command failed or your claims contradict " +
+            "the evidence. You must: (a) run the required command and report its " +
+            "actual output, (b) reconcile your success claim with the evidence — " +
+            "do NOT report success when the evidence contradicts it, and " +
+            "(c) resubmit an honest SelfEvaluation for the SAME round.",
+        check: contradiction.check,
+    });
+}
+/** R3: Agent claims success but did nothing verifiable.
+ *  v1.17: execution_evidence is now MANDATORY for structured self-evaluations.
+ *  Missing evidence when success=true → reject (agent must provide evidence).
+ *  Empty evidence (no files + no tests) when success=true → reject. */
+function enforceEmptySuccess(selfEval, _flags, skipEvidenceRules = false) {
+    if (skipEvidenceRules)
+        return null; // Requires execution_evidence
+    if (!effectiveSuccess(selfEval))
+        return null;
+    const ev = selfEval.execution_evidence;
+    // v1.17: Missing evidence when claiming success → reject.
+    // The agent MUST provide execution_evidence to back up a success claim.
+    if (!ev) {
+        return makeEnforcementResult({
+            action: "reject",
+            reason: "Agent claims success but provided no execution_evidence. " +
+                "Every successful round MUST include execution_evidence with " +
+                "files_changed, test_results, and progress_estimate.",
+            fix_instructions: "You must provide execution_evidence in your self-evaluation: " +
+                "(a) list the files you changed in execution_evidence.files_changed, " +
+                "(b) run tests and report results in execution_evidence.test_results, " +
+                "(c) estimate your progress in execution_evidence.progress_estimate. " +
+                "If you genuinely completed the task without file changes or tests, " +
+                "explain why in detail in your output_summary.",
+            check: "empty_success",
+        });
+    }
+    const filesEmpty = ev.files_changed.length === 0;
+    const testsNotRun = ev.test_results === null;
+    if (!filesEmpty || !testsNotRun)
+        return null;
+    return makeEnforcementResult({
+        action: "reject",
+        reason: "Agent claims success but execution_evidence shows no files changed " +
+            "and no tests were run. There is no verifiable evidence of work.",
+        fix_instructions: "You must provide verifiable evidence: " +
+            "(a) list the files you changed in execution_evidence.files_changed, " +
+            "and (b) run tests and report results in execution_evidence.test_results. " +
+            "If you genuinely completed the task without file changes or tests, " +
+            "explain why in detail in your output_summary.",
+        check: "empty_success",
+    });
+}
+/** R8 (v2.12): Success claimed without any machine-verifiable evidence.
+ *  The verification gate produced success_without_verified_evidence (error):
+ *  zero verified claims, no test results. First occurrence → reject with
+ *  concrete evidence requirements; second consecutive → terminate. */
+function enforceSuccessWithoutVerifiedEvidence(flags, consecutiveRejections) {
+    const flag = flags.find((f) => f.check === CHECK_SUCCESS_WITHOUT_VERIFIED_EVIDENCE && f.severity === "error");
+    if (!flag)
+        return null;
+    if (consecutiveRejections >= 2) {
+        return makeEnforcementResult({
+            action: "terminate",
+            reason: "Repeated success claims with zero machine-verifiable evidence.",
+            check: "success_without_verified_evidence",
+        });
+    }
+    const escalation = getPolicy().engine.enforcement_escalation_enabled &&
+        consecutiveRejections >= 1
+        ? buildEscalationNotice()
+        : "";
+    return makeEnforcementResult({
+        action: "reject",
+        reason: "Agent claims success but the runtime verified zero claims: no passing " +
+            "test evidence and no test results were observed. Success must be backed " +
+            "by machine-verifiable evidence.",
+        fix_instructions: "Provide verifiable evidence for your success claim: (a) run tests and " +
+            "report results in execution_evidence.test_results, or (b) configure a " +
+            "required verification command and run it, or (c) if no change was " +
+            "genuinely needed, declare no_change_reason explaining why." + escalation,
+        check: "success_without_verified_evidence",
+    });
+}
+/** R9 (v2.12): Post-backtrack workspace not restored. The verification gate
+ *  produced backtrack_workspace_not_restored (error): either skipped-round
+ *  files are still dirty, or the working tree is not at the restore point's
+ *  git HEAD. The agent must restore the workspace, not redo the round. */
+function enforceBacktrackNotRestored(flags, vaultEntries = [], currentRound = 0) {
+    const flag = flags.find((f) => f.check === CHECK_BACKTRACK_WORKSPACE_NOT_RESTORED && f.severity === "error");
+    if (!flag)
+        return null;
+    if (!getPolicy().engine.backtrack_enabled)
+        return null;
+    // v2.14: same deadlock guard as R4/R5 — a backtrack that already fired
+    // for this round means a non-restoring agent would otherwise cycle
+    // R9-backtrack forever (backtracks do not increment the rejection
+    // counter). One rollback chance, then terminate.
+    if (currentRound > 0 && hasCommittedBacktrack(vaultEntries, currentRound)) {
+        return makeEnforcementResult({
+            action: "terminate",
+            reason: "The workspace is still not restored after the previous backtrack: " +
+                flag.detail,
+            check: "backtrack_workspace_not_restored",
+        });
+    }
+    return makeEnforcementResult({
+        action: "backtrack",
+        reason: "The workspace was not restored after backtrack: " + flag.detail,
+        fix_instructions: "Restore the working directory to the clean round's state " +
+            "(git checkout -- . && git clean -fd, or git stash; if the failed " +
+            "rounds created commits, git reset --hard <restore-commit>) before " +
+            "submitting the next round.",
+        check: "backtrack_workspace_not_restored",
+    });
+}
+/** R4: Progress has stalled for 3+ consecutive rounds.
+ *  Detected by checking progress_estimate deltas across vault entries.
+ *  First occurrence → REJECT (agent must change approach).
+ *  Second consecutive occurrence → TERMINATE (agent cannot recover). */
+function enforceProgressStall(selfEval, _flags, currentRound, vaultEntries, consecutiveRejections, skipEvidenceRules = false) {
+    // Need at least 3 rounds of history
+    if (currentRound < 3)
+        return null;
+    let isStalling = false;
+    let stallDetail = "";
+    let stallThreshold = 0;
+    if (skipEvidenceRules) {
+        // v3.2: heuristic path has no execution_evidence — fall back to machine
+        // observations: three consecutive rounds with no git change = stall.
+        // No git signal → keep the legacy skip (the machine cannot observe).
+        const machine = machineProgressSeries(vaultEntries, currentRound, 3);
+        if (machine === null)
+            return null;
+        isStalling = machine.every((value) => !value);
+        stallDetail = "no machine-observed changes in the last 3 rounds";
+    }
+    else {
+        const progressByRound = progressWindow(vaultEntries, currentRound, selfEval);
+        if (progressByRound.size < 3)
+            return null;
+        const sortedRounds = [...progressByRound.keys()].sort((a, b) => a - b);
+        const last3 = sortedRounds.slice(-3);
+        // Require all three data points to be within [currentRound-3, currentRound-1].
+        // (collectProgressByRound already filters rnd < currentRound, so the upper
+        // bound is always satisfied — the lower bound is the real continuity guard.)
+        const expectedMin = currentRound - 3;
+        if (last3[0] < expectedMin)
+            return null;
+        const p1 = progressByRound.get(last3[0]);
+        const p2 = progressByRound.get(last3[1]);
+        const p3 = progressByRound.get(last3[2]);
+        stallThreshold = getPolicy().evolution.progress_stall_threshold;
+        const delta12 = p2 - p1;
+        const delta23 = p3 - p2;
+        // Both deltas must be below threshold AND not near completion
+        isStalling = delta12 < stallThreshold && delta23 < stallThreshold && p3 < 0.95;
+        stallDetail = `${(p1 * 100).toFixed(0)}% → ${(p2 * 100).toFixed(0)}% → ${(p3 * 100).toFixed(0)}%`;
+    }
+    if (!isStalling)
+        return null;
+    const escalation = getPolicy().engine.enforcement_escalation_enabled;
+    // v2.7: When escalation is enabled, give one extra rejection round with a
+    // "Seek Human Guidance" notice before terminating. This lets the agent
+    // course-correct with human input rather than being killed immediately.
+    if (consecutiveRejections >= 2) {
+        return makeEnforcementResult({
+            action: "terminate",
+            reason: `Progress stalled for 3+ rounds ` +
+                `(${stallDetail}) ` +
+                `and agent did not resolve after escalation. Terminating loop.`,
+            check: "progress_stall",
+        });
+    }
+    if (consecutiveRejections >= 1 && escalation) {
+        // v2.10: When backtrack is enabled, escalate to backtrack instead of
+        // reject. The agent is rolled back to the last clean round with lessons
+        // injected, rather than redoing the same round with an escalation notice.
+        if (getPolicy().engine.backtrack_enabled) {
+            // v2.14: A backtrack that already committed for this round (the loop
+            // is re-walking rolled-back territory with the same stall) must not
+            // loop forever — terminate instead. Without this, consecutive
+            // rejections reset on every backtrack and the R4 cycle has no exit.
+            if (hasCommittedBacktrack(vaultEntries, currentRound)) {
+                return makeEnforcementResult({
+                    action: "terminate",
+                    reason: `Progress has stalled over the last 3 rounds ` +
+                        `(${stallDetail}) and a backtrack already fired for ` +
+                        `this round. The stall persists after rollback — terminating.`,
+                    check: "progress_stall",
+                });
+            }
+            return makeEnforcementResult({
+                action: "backtrack",
+                reason: `Progress has stalled: ${stallDetail} over the last 3 rounds` +
+                    `${stallThreshold > 0 ? ` (delta < ${(stallThreshold * 100).toFixed(0)}% each round)` : ""}.`,
+                fix_instructions: "Your progress has been flat for 3 rounds. You are being rolled back " +
+                    "to the last clean round. Do NOT repeat the approach that led to the stall. " +
+                    "Choose a different technique or task decomposition.",
+                check: "progress_stall",
+            });
+        }
+        return makeEnforcementResult({
+            action: "reject",
+            reason: `Progress has stalled: ${stallDetail} over the last 3 rounds` +
+                `${stallThreshold > 0 ? ` (delta < ${(stallThreshold * 100).toFixed(0)}% each round)` : ""}.`,
+            fix_instructions: "Your progress has been flat for 3 rounds. You must: " +
+                "(a) explain what is blocking progress, " +
+                "(b) propose a DIFFERENT technique or task decomposition, and " +
+                "(c) set a concrete, verifiable goal for the redo of this round. " +
+                "Do NOT repeat the same approach — it has not moved progress forward." +
+                buildEscalationNotice(),
+            check: "progress_stall",
+        });
+    }
+    if (consecutiveRejections >= 1) {
+        return makeEnforcementResult({
+            action: "terminate",
+            reason: `Progress stalled for 3+ rounds ` +
+                `(${stallDetail}) ` +
+                `and agent did not resolve after previous rejection. Terminating loop.`,
+            check: "progress_stall",
+        });
+    }
+    return makeEnforcementResult({
+        action: "reject",
+        reason: `Progress has stalled: ${stallDetail} over the last 3 rounds` +
+            `${stallThreshold > 0 ? ` (delta < ${(stallThreshold * 100).toFixed(0)}% each round)` : ""}.`,
+        fix_instructions: "Your progress has been flat for 3 rounds. You must: " +
+            "(a) explain what is blocking progress, " +
+            "(b) propose a DIFFERENT technique or task decomposition, and " +
+            "(c) set a concrete, verifiable goal for the redo of this round. " +
+            "Do NOT repeat the same approach — it has not moved progress forward.",
+        check: "progress_stall",
+    });
+}
+/** R5: Progress completely stalled (delta = 0) for N consecutive rounds.
+ *  Stricter than R4 (which triggers on delta < 5%). R4 gives the agent a chance
+ *  to change approach via reject; this rule fires when progress has been exactly
+ *  flat — the agent is making zero forward motion.
+ *
+ *  v2.7: When enforcement_escalation_enabled is true, the first occurrence
+ *  issues a rejection with an escalation notice instead of terminating.
+ *  Termination only happens on the second consecutive flatline detection.
+ *
+ *  Uses the engine.max_circuit_breaker policy value as the lookback window. */
+function enforceProgressStallTerminal(selfEval, _flags, currentRound, vaultEntries, consecutiveRejections, skipEvidenceRules = false) {
+    const breakerSize = getPolicy().engine.max_circuit_breaker;
+    if (currentRound < breakerSize || vaultEntries.length < breakerSize)
+        return null;
+    let allFlat = false;
+    let flatDetail = "";
+    if (skipEvidenceRules) {
+        // v3.2: heuristic path has no execution_evidence — fall back to machine
+        // observations: N consecutive rounds with no git change = flatline.
+        // No git signal → keep the legacy skip (the machine cannot observe).
+        const machine = machineProgressSeries(vaultEntries, currentRound, breakerSize);
+        if (machine === null)
+            return null;
+        allFlat = machine.every((value) => !value);
+        flatDetail = `no machine-observed changes for ${breakerSize} consecutive rounds`;
+    }
+    else {
+        const progressByRound = progressWindow(vaultEntries, currentRound, selfEval);
+        if (progressByRound.size < breakerSize)
+            return null;
+        const sortedRounds = [...progressByRound.keys()].sort((a, b) => a - b);
+        const recent = sortedRounds.slice(-breakerSize);
+        // v2.14: continuity guard (mirrors R4) — "consecutive" means the most
+        // recent rounds. Rounds without execution_evidence are unknown motion,
+        // not zero motion; taking "the last N rounds that happened to report
+        // progress" terminated loops whose recent rounds simply had no data.
+        if (recent[0] < currentRound - breakerSize)
+            return null;
+        // Check that progress values are ALL identical (delta = 0).
+        // Use epsilon comparison — JSON round-trips can produce float drift
+        // (e.g. 0.3 vs 0.30000000000000004) that would defeat strict === .
+        const EPSILON = 1e-10;
+        const first = progressByRound.get(recent[0]);
+        allFlat = recent.every((r) => Math.abs((progressByRound.get(r) ?? 0) - first) < EPSILON);
+        if (!allFlat || first >= 0.95)
+            return null;
+        flatDetail = `completely flat at ${(first * 100).toFixed(0)}%`;
+    }
+    if (!allFlat)
+        return null;
+    const escalation = getPolicy().engine.enforcement_escalation_enabled;
+    // v2.7: When escalation is enabled, give one rejection round with a
+    // "Seek Human Guidance" notice before terminating.
+    if (consecutiveRejections >= 1) {
+        return makeEnforcementResult({
+            action: "terminate",
+            reason: `Progress has been ${flatDetail} ` +
+                `for ${breakerSize} consecutive rounds and agent did not resolve ` +
+                `after escalation. The agent is making zero forward motion.`,
+            check: "progress_stall_terminal",
+        });
+    }
+    if (escalation) {
+        // v2.10: When backtrack is enabled, escalate to backtrack instead of
+        // reject. The agent is rolled back to the last clean round with a
+        // "radically different strategy" lesson injected.
+        if (getPolicy().engine.backtrack_enabled) {
+            // v2.14: Same deadlock guard as R4 — a backtrack that already fired
+            // for this round means the loop is re-walking rolled-back territory.
+            if (hasCommittedBacktrack(vaultEntries, currentRound)) {
+                return makeEnforcementResult({
+                    action: "terminate",
+                    reason: `Progress has been ${flatDetail} ` +
+                        `for ${breakerSize} consecutive rounds and a backtrack already ` +
+                        `fired for this round. Flatline persists after rollback — ` +
+                        `terminating the loop.`,
+                    check: "progress_stall_terminal",
+                });
+            }
+            return makeEnforcementResult({
+                action: "backtrack",
+                reason: `Progress has been ${flatDetail} ` +
+                    `for ${breakerSize} consecutive rounds. The agent is making zero ` +
+                    `forward motion.`,
+                fix_instructions: "Your progress has been exactly flat — zero forward motion. " +
+                    "You are being rolled back to the last clean round. " +
+                    "Choose a RADICALLY different approach. The previous one produced nothing.",
+                check: "progress_stall_terminal",
+            });
+        }
+        return makeEnforcementResult({
+            action: "reject",
+            reason: `Progress has been ${flatDetail} ` +
+                `for ${breakerSize} consecutive rounds. The agent is making zero ` +
+                `forward motion.`,
+            fix_instructions: "Your progress has been exactly flat — you are making zero forward " +
+                "motion. You must: (a) explain what is fundamentally blocking you, " +
+                "(b) propose a radically different approach, and " +
+                "(c) set a concrete verifiable goal. " +
+                "Do NOT resubmit the same output." +
+                buildEscalationNotice(),
+            check: "progress_stall_terminal",
+        });
+    }
+    return makeEnforcementResult({
+        action: "terminate",
+        reason: `Progress has been ${flatDetail} ` +
+            `for ${breakerSize} consecutive rounds. The agent is making zero ` +
+            `forward motion and cannot recover.`,
+        check: "progress_stall_terminal",
+    });
+}
+/** R6: Two consecutive rejections → escalate (v2.7) or terminate.
+ *
+ *  v2.7: When enforcement_escalation_enabled is true, the second consecutive
+ *  rejection issues an escalated rejection with a "Seek Human Guidance" notice
+ *  instead of terminating immediately. Termination occurs on the third
+ *  consecutive rejection. When the flag is disabled, behavior is unchanged
+ *  (terminate at ≥2 rejections). */
+function enforceMaxRejections(consecutiveRejections) {
+    const escalation = getPolicy().engine.enforcement_escalation_enabled;
+    if (escalation) {
+        if (consecutiveRejections >= 3) {
+            return makeEnforcementResult({
+                action: "terminate",
+                reason: `${consecutiveRejections} consecutive enforcement rejections for the ` +
+                    `same issue without resolution. The agent has been unable to correct ` +
+                    `the identified issue even after escalation. Terminating loop.`,
+                check: "max_rejections",
+            });
+        }
+        if (consecutiveRejections >= 2) {
+            return makeEnforcementResult({
+                action: "reject",
+                reason: `${consecutiveRejections} consecutive enforcement rejections for the ` +
+                    `same issue without resolution.`,
+                fix_instructions: "You have been rejected multiple times for the same issue. " +
+                    "You must address ALL items in the Evidence Gap section above. " +
+                    "If you do not understand why you are being rejected, explain " +
+                    "your confusion explicitly and ask for clarification." +
+                    buildEscalationNotice(),
+                check: "max_rejections",
+            });
+        }
+        return null;
+    }
+    if (consecutiveRejections < 2)
+        return null;
+    return makeEnforcementResult({
+        action: "terminate",
+        reason: `${consecutiveRejections} consecutive enforcement rejections for the ` +
+            `same issue without resolution. The agent has been unable to correct the identified issue.`,
+        check: "max_rejections",
+    });
+}
+/** v2.12: Check if a drift_clarification contains at least one semantic
+ *  anchor — a constraint/criterion/sub-goal ID (v2.11) or a file-path-like
+ *  reference. This distinguishes genuine explanations from filler text that
+ *  just happens to be ≥ 20 characters. */
+/** Extract anchor candidates from a clarification: c-/cr-/sg- IDs and
+ *  file-path-like tokens. */
+function extractAnchors(clarification) {
+    const anchors = new Set();
+    for (const match of clarification.matchAll(/[c]r?-[a-f0-9]{8}|sg-[a-f0-9]{8}/gi)) {
+        anchors.add(match[0].toLowerCase());
+    }
+    for (const match of clarification.matchAll(/[\w./-]+\.[a-z]{2,6}\b/gi)) {
+        anchors.add(match[0]);
+    }
+    return [...anchors];
+}
+/** v2.12: A drift clarification is substantive only when at least one anchor
+ *  is REAL — a stable ID that derives from a known constraint/criterion/
+ *  sub-goal text in the vault, or a file path present in reported/observed
+ *  files. Regex patterns alone (e.g. a fabricated `c-00000000`) no longer
+ *  count, closing the v2.12 filler-text loophole. */
+function clarificationIsSubstantive(clarification, selfEval, vaultEntries) {
+    const anchors = extractAnchors(clarification);
+    if (anchors.length === 0)
+        return false;
+    const knownFiles = new Set();
+    const knownTexts = new Set();
+    const collect = (values) => {
+        if (!Array.isArray(values))
+            return;
+        for (const value of values) {
+            if (typeof value === "string" && value.length > 0)
+                knownTexts.add(value);
+        }
+    };
+    // The agent's own round report is a known set: its constraints, criteria,
+    // and files are legitimate anchor targets.
+    for (const file of selfEval.execution_evidence?.files_changed ?? [])
+        knownFiles.add(file);
+    collect(selfEval.discovered_constraints);
+    collect(selfEval.constraint_violations);
+    const selfEvidence = selfEval.execution_evidence;
+    if (selfEvidence) {
+        collect(selfEvidence.success_criteria_met);
+        collect(selfEvidence.success_criteria_remaining);
+    }
+    for (const entry of vaultEntries) {
+        collect(entry.constraint_violations);
+        collect(entry.discovered_constraints);
+        const ev = entry.execution_evidence;
+        if (ev) {
+            collect(ev.files_changed);
+            collect(ev.success_criteria_met);
+            collect(ev.success_criteria_remaining);
+        }
+    }
+    for (const anchor of anchors) {
+        if (STABLE_ID_RE.test(anchor)) {
+            const matches = anchor.startsWith("c-") && !anchor.startsWith("cr-")
+                ? [...knownTexts].some((text) => deriveConstraintId(text) === anchor)
+                : anchor.startsWith("cr-")
+                    ? [...knownTexts].some((text) => deriveCriterionId(text) === anchor)
+                    : [...knownTexts].some((text) => deriveSubGoalId(text) === anchor);
+            if (matches)
+                return true;
+            continue;
+        }
+        // File path: exact or suffix match against known files.
+        if ([...knownFiles].some((file) => file === anchor || file.endsWith(`/${anchor}`) || file.endsWith(`\\${anchor}`))) {
+            return true;
+        }
+    }
+    return false;
+}
+/** Build a stronger rejection notice as the clarification streak grows. */
+function buildClarificationRejectionNotice(streak) {
+    const base = "Your declared next_action from the previous round does not match " +
+        "what you actually did this round. You must either: " +
+        "(a) explain why you pivoted in the drift_clarification field, " +
+        "mentioning specific constraint/criterion/sub-goal IDs or file paths " +
+        "that prompted the change, or " +
+        "(b) redo this round and do what you said you would do.";
+    if (streak >= 2) {
+        return (base +
+            "\n\n### ⚠️ Repeated Weak Clarification\n\n" +
+            `You have submitted drift_clarification without concrete references ` +
+            `for ${streak} consecutive rounds. Your explanation MUST include at ` +
+            `least one of: a constraint ID (c-XXXXXXXX), criterion ID (cr-XXXXXXXX), ` +
+            `sub-goal ID (sg-XXXXXXXX), or a file path you changed.\n\n` +
+            `🚨 If the next clarification also lacks concrete anchors, the loop ` +
+            `will be terminated.`);
+    }
+    return base;
+}
+/** R7: Agent declared a next_action but did something unrelated.
+ *
+ *  v2.8: When the agent provides a substantive drift_clarification (≥ 20
+ *  characters), the rejection is waived — the agent has acknowledged and
+ *  explained the pivot. An empty or absent clarification still triggers
+ *  the normal reject/terminate path.
+ *
+ *  v2.12: Clarification now requires a semantic anchor (constraint/criterion/
+ *  sub-goal ID or file path) in addition to the ≥ 20 char minimum. Weak
+ *  clarifications (no anchors) increment a streak counter; 3 consecutive
+ *  weak clarifications terminate the loop.
+ *
+ *  First occurrence → reject (agent must explain the gap or realign).
+ *  Second consecutive → terminate (agent cannot/will not follow its own plan). */
+function enforceIntentDrift(flags, selfEval, consecutiveRejections, driftClarificationStreak = 0, vaultEntries = []) {
+    const flag = flags.find((f) => f.check === "intent_drift");
+    if (!flag)
+        return null;
+    const clarification = selfEval.drift_clarification?.trim();
+    // v2.12: Substantive anchors — the referenced ID/path must actually exist
+    // in the vault or in reported files; a fabricated ID no longer counts.
+    const hasAnchor = clarification
+        ? clarificationIsSubstantive(clarification, selfEval, vaultEntries)
+        : false;
+    // v2.12: Substantive clarification = length check + semantic anchor
+    const isSubstantive = clarification
+        && clarification.length >= 20
+        && hasAnchor;
+    if (isSubstantive) {
+        // Genuine explanation with concrete references → accept the pivot
+        return makeEnforcementResult({
+            action: "accept",
+            reason: "",
+            fix_instructions: "",
+            check: "intent_drift",
+            clarification_accepted: true,
+        });
+    }
+    // v2.12: Clarification present but weak (no semantic anchors)
+    if (clarification && clarification.length >= 20 && !hasAnchor) {
+        const maxStreak = getPolicy().engine.drift_clarification_max_streak;
+        // When maxStreak is 0, disable the check (pre-v2.12 behavior)
+        if (maxStreak <= 0) {
+            return makeEnforcementResult({
+                action: "accept",
+                reason: "",
+                fix_instructions: "",
+                check: "intent_drift",
+                clarification_accepted: true,
+            });
+        }
+        const newStreak = driftClarificationStreak + 1;
+        if (newStreak >= maxStreak) {
+            return makeEnforcementResult({
+                action: "terminate",
+                reason: `Agent has submitted weak drift_clarification (no concrete IDs or ` +
+                    `file references) for ${newStreak} consecutive rounds. The agent ` +
+                    `systematically avoids following its own plan without substantive ` +
+                    `explanation.`,
+                check: "intent_drift",
+            });
+        }
+        return makeEnforcementResult({
+            action: "reject",
+            reason: flag.detail,
+            fix_instructions: buildClarificationRejectionNotice(newStreak),
+            check: "intent_drift",
+        });
+    }
+    // No clarification or too short — the worst form of drift. v2.14: counts
+    // toward the SAME streak as weak clarifications (maxStreak consecutive
+    // un-explained drifts terminate, per the README's "three consecutive weak
+    // clarifications"). Previously this branch used the GLOBAL
+    // consecutiveRejections counter, so an unrelated earlier rejection
+    // (e.g. R1) terminated the loop on the very first drift.
+    const maxStreak = getPolicy().engine.drift_clarification_max_streak;
+    // When maxStreak is 0, disable the check (pre-v2.12 behavior)
+    if (maxStreak <= 0) {
+        return makeEnforcementResult({
+            action: "accept",
+            reason: "",
+            fix_instructions: "",
+            check: "intent_drift",
+            clarification_accepted: true,
+        });
+    }
+    const newStreak = driftClarificationStreak + 1;
+    if (newStreak >= maxStreak) {
+        return makeEnforcementResult({
+            action: "terminate",
+            reason: `Agent has drifted from its declared next_action for ${newStreak} ` +
+                `consecutive rounds without any drift_clarification. The agent ` +
+                `cannot or will not follow its own plan — terminating loop.`,
+            check: "intent_drift",
+        });
+    }
+    return makeEnforcementResult({
+        action: "reject",
+        reason: flag.detail,
+        fix_instructions: "Your declared next_action from the previous round does not match " +
+            "what you actually did this round. You must either: " +
+            "(a) explain why you pivoted and what you learned in the " +
+            "drift_clarification field, or " +
+            "(b) redo this round and do what you said you would do. " +
+            "If you intentionally changed direction, update your objective_refinement " +
+            "and set a new next_action that reflects the new plan.",
+        check: "intent_drift",
+    });
+}
+// ═══════════════════════════════════════════════════════════════════════════
+// Escalation notice builder (v2.7)
+// ═══════════════════════════════════════════════════════════════════════════
+/** Build an escalation notice urging the agent to seek human guidance.
+ *
+ *  Appended to fix_instructions when a rule would otherwise terminate.
+ *  Gives the agent one final chance to correct course with human input
+ *  before the loop is irreversibly stopped. */
+function buildEscalationNotice() {
+    return [
+        "",
+        "### 🆘 Escalation — Seek Human Guidance",
+        "",
+        "You have been rejected multiple times for the same issue. " +
+            "Before retrying, pause and ask a human for help:",
+        "",
+        "1. Explain what you've tried so far and why each approach failed.",
+        "2. Ask the human for specific guidance on what to try next.",
+        "3. Incorporate their feedback before re-submitting your self-evaluation.",
+        "",
+        "⚠️ If this round is rejected again, the loop will be **terminated**. " +
+            "Do NOT resubmit without meaningful changes.",
+    ].join("\n");
+}
+// ═══════════════════════════════════════════════════════════════════════════
+// Diagnostic gap builder (v2.7)
+// ═══════════════════════════════════════════════════════════════════════════
+/** Build a diagnostic gap section from verification flags.
+ *
+ *  Translates each error/warn-level VerificationFlag into a concrete
+ *  claim-vs-evidence mismatch statement. Info-level flags are excluded
+ *  because they are purely informational and do not represent gaps.
+ *
+ *  Returns an empty string when there are no diagnostic flags to report. */
+function buildDiagnosticGap(flags) {
+    const diagnosticFlags = flags.filter((f) => f.severity === "error" || f.severity === "warn");
+    if (diagnosticFlags.length === 0)
+        return "";
+    const lines = [
+        "### Evidence Gap",
+        "",
+        "The enforcement gate detected mismatches between your claims and the evidence. " +
+            "Each item below shows a specific discrepancy — address every one before re-submitting.",
+        "",
+    ];
+    for (const flag of diagnosticFlags) {
+        const icon = flag.severity === "error" ? "🚫" : "⚠️";
+        lines.push(`- ${icon} **[${flag.check}]** \`${flag.field}\` — ${flag.detail}`);
+    }
+    lines.push("");
+    return lines.join("\n");
+}
+// ═══════════════════════════════════════════════════════════════════════════
+// Main entry point
+// ═══════════════════════════════════════════════════════════════════════════
+/** Enforce round-boundary rules based on the verification gate's findings
+ *  and the agent's self-evaluation integrity.
+ *
+ *  Rules run in priority order (R1 → R5). The first rule that fires wins.
+ *
+ * @param selfEval              The agent's self-evaluation for the current round.
+ * @param verifyResult          The verification gate's output (from verifySelfEvaluation).
+ * @param currentRound          Current round number (1-based, BEFORE increment).
+ * @param vaultEntries          Vault entries for this loop (used for progress tracking).
+ * @param consecutiveRejections How many consecutive rounds have already been rejected.
+ *                              Starts at 0; increments on each reject; resets on accept.
+ */
+export function enforceRound(selfEval, verifyResult, currentRound, vaultEntries, consecutiveRejections = 0, skipEvidenceRules = false, 
+/** v2.12: Current clarification streak for R7 escalation. */
+driftClarificationStreak = 0) {
+    const { flags } = verifyResult;
+    // Run enforcement rules in priority order.
+    // Earlier rules take higher precedence.
+    const rules = [
+        () => enforceSuccessWithRemainingCriteria(flags),
+        () => enforceRecurringViolation(flags),
+        () => enforceEmptySuccess(selfEval, flags, skipEvidenceRules),
+        // v2.14: R-EVID — a required command failure or self-contradictory
+        // outcome claim contradicts the success claim before commit
+        () => enforceEvidenceContradiction(flags),
+        // v2.12: R8 — success with zero machine-verifiable evidence
+        () => enforceSuccessWithoutVerifiedEvidence(flags, consecutiveRejections),
+        () => enforceProgressStall(selfEval, flags, currentRound, vaultEntries, consecutiveRejections, skipEvidenceRules),
+        () => enforceProgressStallTerminal(selfEval, flags, currentRound, vaultEntries, consecutiveRejections, skipEvidenceRules),
+        () => enforceMaxRejections(consecutiveRejections),
+        () => enforceIntentDrift(flags, selfEval, consecutiveRejections, driftClarificationStreak, vaultEntries),
+        // v2.12: R9 — post-backtrack workspace not restored
+        () => enforceBacktrackNotRestored(flags, vaultEntries, currentRound),
+    ];
+    for (const rule of rules) {
+        const result = rule();
+        if (result)
+            return result;
+    }
+    // All rules passed — accept the round
+    return makeEnforcementResult({
+        action: "accept",
+        reason: "",
+        fix_instructions: "",
+    });
+}
+/** Build a rejection prompt for the agent.
+ *
+ *  The prompt clearly states the round was rejected, why, what the agent
+ *  must fix, and that the agent must redo the SAME round (not advance).
+ *
+ *  v2.7: Accepts verificationFlags to render a diagnostic "Evidence Gap"
+ *  section. Each error/warn flag becomes a concrete claim-vs-evidence
+ *  mismatch statement so the agent knows exactly what to correct rather
+ *  than retrying blindly.
+ *
+ * @param currentRound      The round number that was rejected (NOT incremented).
+ * @param task              The original loop task description.
+ * @param enforceResult     The enforcement decision with reason and fix instructions.
+ * @param verificationFlags The verification gate's findings for this round.
+ *                          Used to build the Evidence Gap section. */
+export function buildRejectionPrompt(currentRound, task, enforceResult, verificationFlags = []) {
+    const lines = [
+        "## ⛔ Round " + currentRound + " — REJECTED",
+        "",
+        "Your self-evaluation for Round " + currentRound +
+            " was **rejected** by the enforcement gate.",
+        "",
+        "### Reason",
+        "",
+        enforceResult.reason,
+        "",
+    ];
+    // v2.7: Diagnostic gap — show the agent exactly what doesn't match.
+    const gap = buildDiagnosticGap(verificationFlags);
+    if (gap) {
+        lines.push(gap);
+    }
+    if (enforceResult.fix_instructions) {
+        lines.push("### Required Fix");
+        lines.push("");
+        lines.push(enforceResult.fix_instructions);
+        lines.push("");
+    }
+    lines.push("### Your Task (Round " + currentRound + " — Retry)");
+    lines.push("");
+    lines.push(task);
+    lines.push("");
+    lines.push("### Instructions");
+    lines.push("");
+    lines.push("1. Read and address each issue in **Required Fix** above.", "2. Re-execute **Round " + currentRound +
+        "** — do NOT advance to the next round.", "3. Submit a corrected self-evaluation via `loopforge_next`.", "4. Be honest in your self-evaluation — " +
+        "claiming success when criteria are unmet will be rejected again.", "5. If you believe this rejection is incorrect, " +
+        "explain why in your output_summary and the enforcement gate will re-evaluate.");
+    return lines.join("\n");
+}
+//# sourceMappingURL=enforcement-gate.js.map