@codyswann/lisa 2.255.0 → 2.256.1

package/plugins/lisa-cursor/rules/config-resolution.mdc

@@ -53,13 +53,32 @@ empirical evidence only—native Playwright/Cypress/Maestro remains the regressi
 
 ## Env → base branch
 
-For implementation work, map the work item's `## Target Backend Environment` to
-the PR base branch through `.lisa.config.json` `deploy.branches`. For bugs, a
-reported environment extracted from the description/reproduction steps wins over
-an autofill default. If a reported environment is absent from `deploy.branches`,
-stop and report the missing mapping instead of silently falling back to the
-default or integration branch. A non-integration environment bug is fixed,
-merged, and verified on that environment branch first, then forward
-cherry-picked down to the integration branch via a linked follow-up.
+The `## Target Backend Environment` grammar is durable: human-confirmed values
+are a bare configured key or `Confirmed: <env>`; automation writes either
+`Inferred: <env> — evidence: <title|body|reproduction|hostname>` or
+`Assumption: <env> — remote default branch <branch>`. When no unique environment
+maps to the default branch, use `Assumption: remote default branch <branch>`.
+Human confirmation replaces an automated annotation with a bare key or
+`Confirmed: <env>`. For legacy bare values, use managed draft markers and current
+ticket content only; provider edit history is not required. A marker proves
+automation and requires re-annotation; otherwise unknown provenance with
+conflicting evidence stops for confirmation.
+
+Resolution order is human-confirmed value; validated `Inferred:` evidence; then
+one unambiguous signal from the human-authored title, body, and reproduction
+steps or a URL hostname. Exclude the complete `Target Backend Environment`
+section and other machine-authored metadata/draft blocks from the scan so an
+annotation cannot validate or conflict with itself. Evidence supersedes only an
+`Assumption:`. A signal must be an exact
+`deploy.branches` key as a complete token/hostname label. The sole normalization
+is `prod` ↔ `production` when exactly one is configured; no other aliases exist.
+Conflicting signals stop. Never infer from arbitrary branch text, URL paths or
+query strings, or substrings. With no signals, use the remote default branch and
+record an assumption: include `<env>` only for a unique reverse-map, otherwise
+use the branch-only form without inventing an environment or blocking. Require
+any selected environment mapping and its remote branch to exist.
+A non-integration environment bug is fixed, merged, and verified on that
+environment branch first, then
+forward cherry-picked down to the integration branch via a linked follow-up.
 
 Full reference: [reference/config-resolution.md](config-resolution-reference.mdc).

package/plugins/lisa-cursor/rules/pre-flight-autofill-reference.mdc

@@ -54,14 +54,29 @@ spec.
 - **Validation Journey (draft)** — run the vendor `*-add-journey` skill to draft
   the click-path, markers, and viewports from the reproduction steps. This is a
   draft for human approval, not the final ratified contract.
-- **Target Backend Environment** — for bugs, first parse the reported
-  environment from the title, description/body, screenshots captions, links, and
-  reproduction steps. Bare words such as `dev`, `staging`, `prod`, and
-  `production` count, as do environment-bearing URLs such as
-  `staging.<domain>`, `gql.staging.*`, and `dev.<domain>`. That reported
-  environment is authoritative for the draft. Only recommend the repo's
-  remote default branch when no environment is discoverable anywhere in the
-  work item, and record the default as an explicit assumption.
+- **Target Backend Environment** — preserve a bare configured key or
+  `Confirmed: <env>` as human-confirmed. Automated evidence must write
+  `Inferred: <env> — evidence: <title|body|reproduction|hostname>`; a generic
+  fallback must write `Assumption: <env> — remote default branch <branch>`.
+  If the default branch has no unique environment reverse-map, instead write
+  `Assumption: remote default branch <branch>` without inventing an environment.
+  Human confirmation replaces an automated annotation with a bare key or
+  `Confirmed: <env>`. For legacy bare values, use managed draft markers and
+  current ticket content only; provider edit history is not required. A marker
+  proves automation and requires re-annotation; otherwise unknown provenance
+  plus conflicting evidence stops for confirmation. A human-confirmed value wins, followed by validated
+  `Inferred:` evidence. Otherwise inspect the title, body, and reproduction
+  steps or a URL hostname for one unambiguous exact `deploy.branches` key,
+  excluding the complete `Target Backend Environment` section and all other
+  machine-authored metadata/draft blocks so annotations cannot become evidence.
+  Clear evidence supersedes only an `Assumption:` value. The sole normalization
+  is `prod` ↔ `production` when exactly one key is configured; no other aliases
+  exist. Never infer from arbitrary branch text, URL paths or query strings, or
+  substrings. Multiple conflicting signals **stop** autofill. With no
+  environment signal, use the remote default and record the applicable
+  `Assumption:` form; a missing/ambiguous reverse-map alone never blocks.
+  Require any selected environment mapping and remote branch. Never overwrite
+  human prose.
 
 **Tier B — irreducibly human (cannot invent — but still propose a default):**
 

package/plugins/lisa-cursor/rules/pre-flight-autofill.mdc

@@ -26,13 +26,27 @@ information was already on the ticket; only the structure was missing).
   Gherkin Acceptance Criteria (one scenario per fix), expected-vs-actual +
   environment, Repository, Relationship Search (actually run the git+tracker
   search — don't fabricate "none found"), Validation Journey **draft** (via the
-  vendor `*-add-journey` skill), Target Backend Environment. For bugs, parse the
-  reported environment from the item body/reproduction steps before recommending
-  any default: bare environment names (`dev`, `staging`, `prod`, `production`)
-  and environment-bearing URLs (for example `staging.<domain>`, `gql.staging.*`,
-  `dev.<domain>`) are authoritative. Only recommend the repo's remote default branch when
-  no environment is discoverable anywhere in the work item, and record that as
-  an assumption.
+  vendor `*-add-journey` skill), Target Backend Environment. Preserve a bare
+  configured key or `Confirmed: <env>` as human-confirmed. Automation writes
+  `Inferred: <env> — evidence: <title|body|reproduction|hostname>` or
+  `Assumption: <env> — remote default branch <branch>`. Without a unique
+  reverse-map use `Assumption: remote default branch <branch>`; human confirmation
+  replaces the annotation with a bare key or `Confirmed: <env>`. For legacy bare
+  values, use managed draft markers and current ticket content only; provider
+  edit history is not required. A marker proves automation and requires
+  re-annotation; otherwise unknown provenance plus conflicting evidence stops.
+  Human-confirmed wins, then validated `Inferred:`
+  evidence. Otherwise use one unambiguous exact `deploy.branches` key from the
+  title, body, and reproduction steps or a URL hostname, excluding the complete
+  `Target Backend Environment` section and other machine-authored metadata/draft
+  blocks so annotations cannot become evidence. Evidence supersedes only
+  `Assumption:`. Normalize only built-in
+  `prod` ↔ `production` when exactly one is configured; no other aliases exist.
+  Never infer from arbitrary branch text, URL paths/query strings, or substrings.
+  Conflicting signals **stop** autofill. With no environment signal, use the
+  remote default and record the applicable `Assumption:` form; a non-unique
+  reverse-map alone never blocks. Require any selected mapping and remote branch.
+  Never overwrite human prose.
 - **Tier B — irreducibly human:** real credentials/access that exist nowhere on
   the item or in repo test-user docs; a genuine product/scoping decision with no
   defensible default. Still propose a recommended default where one exists; ask

package/plugins/lisa-cursor/scripts/automation-status-claude-adapter.mjs

@@ -15,6 +15,10 @@ import {
   createAutomationGroupBins,
   renderAutomationGroups,
 } from "./automation-status-expected-fleet.mjs";
+import {
+  resolveAutomationRunDisplay,
+  resolveRecoveryEscalation,
+} from "./automation-status-run-history.mjs";
 
 const CLAUDE_RUNTIME_LABEL = "Claude /schedule";
 const CLAUDE_ACTIVE_STATUSES = new Set([
@@ -51,9 +55,10 @@ const NEGATED_FAILURE_PATTERN =
  * @param {{
  *   readonly expectedFleet: ExpectedFleet
  *   readonly scheduleListing?: string | readonly unknown[] | Record<string, unknown> | null
+ *   readonly projectRoot?: string
  *   readonly now?: string | Date
  * }} input
- * @returns {{
+ * @returns {Promise<{
  *   readonly runtime: string
  *   readonly generatedAt: string
  *   readonly groups: readonly {
@@ -66,20 +71,27 @@ const NEGATED_FAILURE_PATTERN =
  *       readonly expectedCadence?: string
  *       readonly expectedCommand?: string
  *       readonly observed?: string
+ *       readonly runbook?: string
+ *       readonly lastOutcome?: { readonly ts: string, readonly outcome: string, readonly summary: string }
+ *       readonly outcomeHistory?: readonly string[]
+ *       readonly olderRecordCount?: number
  *       readonly remediation?: string
  *     }[]
  *   }[]
  *   readonly observedAutomations: readonly ObservedClaudeAutomation[]
- * }}}
+ * }>}
  */
-export function inspectClaudeAutomationFleet(input) {
+export async function inspectClaudeAutomationFleet(input) {
   const expectedFleet = input.expectedFleet;
   const now = normalizeDate(input.now);
+  const projectRoot = input.projectRoot ?? process.cwd();
   const observedAutomations = listClaudeAutomations({
     scheduleListing: input.scheduleListing,
     automationPrefix: expectedFleet.automationPrefix,
   });
 
+  const runDisplays = await resolveFleetRunDisplays(projectRoot, expectedFleet);
+
   const expectedGroups = createAutomationGroupBins();
 
   const comparisons = compareAutomationFleet({
@@ -96,17 +108,23 @@ export function inspectClaudeAutomationFleet(input) {
         expected,
         comparison,
         now,
+        runDisplay: runDisplays.get(expected.id),
       })
     );
   }
 
   for (const unsupported of expectedFleet.unsupported) {
+    const runDisplay = runDisplays.get(unsupported.id);
     assignToAutomationGroup(expectedGroups, unsupported.group, {
       id: unsupported.automationId,
       status: "UNSUPPORTED",
       summary: unsupported.reason,
       expectedCadence: unsupported.expectedCadence,
       observed: "No automation is expected for this repo/runtime combination.",
+      runbook: runDisplay?.runbook,
+      lastOutcome: runDisplay?.lastOutcome,
+      outcomeHistory: runDisplay?.outcomeHistory,
+      olderRecordCount: runDisplay?.olderRecordCount,
     });
   }
 
@@ -195,15 +213,43 @@ function extractClaudeScheduleCadence(command) {
   );
 }
 
+/**
+ * Read the read-only run-history display for every expected and unsupported
+ * loop up front, keyed by the loop's short id, so the per-entry item builder
+ * stays synchronous.
+ *
+ * @param {string} projectRoot
+ * @param {ExpectedFleet} expectedFleet
+ * @returns {Promise<Map<string, import("./automation-status-run-history.mjs").AutomationRunDisplay>>}
+ */
+async function resolveFleetRunDisplays(projectRoot, expectedFleet) {
+  const entries = [...expectedFleet.expected, ...expectedFleet.unsupported];
+  const displays = await Promise.all(
+    entries.map(entry =>
+      resolveAutomationRunDisplay({
+        projectRoot,
+        loopId: entry.id,
+        runbookPath: entry.runbookPath,
+      })
+    )
+  );
+  return new Map(entries.map((entry, index) => [entry.id, displays[index]]));
+}
+
 function createObservedStatusItem(input) {
   const expected = input.expected;
   const comparison = input.comparison;
   const observed = comparison.observedAutomation;
+  const runDisplay = input.runDisplay;
   const runSignal = classifyAutomationRunSignal({
     expected,
     observedAutomation: observed,
     now: input.now,
   });
+  // Three or more consecutive recorded recovery-required runs flip the loop to
+  // FAILING even when the /schedule entry looks fine — the fleet-health signal
+  // the scheduler surface cannot see. It wins over the scheduler run-signal.
+  const escalation = resolveRecoveryEscalation(runDisplay);
 
   const observedDetails = [comparison.observed];
   if (observed?.status) {
@@ -223,20 +269,30 @@ function createObservedStatusItem(input) {
   }
 
   const status =
+    escalation?.status ??
     runSignal?.status ??
     /** @type {"HEALTHY" | "MISSING" | "DRIFTED"} */ (comparison.status);
 
   return {
     id: expected.automationId,
     status,
-    summary: composeAutomationSummary({
-      comparison,
-      runSignal,
-    }),
+    summary: escalation
+      ? escalation.summary
+      : composeAutomationSummary({
+          comparison,
+          runSignal,
+        }),
     expectedCadence: expected.expectedCadence,
     expectedCommand: expected.expectedCommand,
     observed: observedDetails.join(" "),
-    remediation: runSignal?.remediation ?? comparison.remediation,
+    runbook: runDisplay?.runbook,
+    lastOutcome: runDisplay?.lastOutcome,
+    outcomeHistory: runDisplay?.outcomeHistory,
+    olderRecordCount: runDisplay?.olderRecordCount,
+    remediation:
+      escalation?.remediation ??
+      runSignal?.remediation ??
+      comparison.remediation,
   };
 }
 

package/plugins/lisa-cursor/scripts/automation-status-codex-adapter.mjs

@@ -22,6 +22,10 @@ import {
   createAutomationGroupBins,
   renderAutomationGroups,
 } from "./automation-status-expected-fleet.mjs";
+import {
+  resolveAutomationRunDisplay,
+  resolveRecoveryEscalation,
+} from "./automation-status-run-history.mjs";
 
 const CODEx_RUNTIME_LABEL = "Codex automations";
 const RUN_TIMESTAMP_PATTERN = /20\d{2}-\d\d-\d\dT\d\d:\d\d:\d\d(?:\.\d+)?Z/;
@@ -132,6 +136,7 @@ async function execGitWithRetry(args, options, attempts = 4) {
  * @param {{
  *   readonly expectedFleet: ExpectedFleet
  *   readonly automationsDir?: string
+ *   readonly projectRoot?: string
  *   readonly now?: string | Date
  * }} input
  * @returns {Promise<{
@@ -147,6 +152,10 @@ async function execGitWithRetry(args, options, attempts = 4) {
  *       readonly expectedCadence?: string
  *       readonly expectedCommand?: string
  *       readonly observed?: string
+ *       readonly runbook?: string
+ *       readonly lastOutcome?: { readonly ts: string, readonly outcome: string, readonly summary: string }
+ *       readonly outcomeHistory?: readonly string[]
+ *       readonly olderRecordCount?: number
  *       readonly remediation?: string
  *     }[]
  *   }[]
@@ -156,11 +165,14 @@ async function execGitWithRetry(args, options, attempts = 4) {
 export async function inspectCodexAutomationFleet(input) {
   const expectedFleet = input.expectedFleet;
   const now = normalizeDate(input.now);
+  const projectRoot = input.projectRoot ?? process.cwd();
   const observedAutomations = await listCodexAutomations({
     automationsDir: input.automationsDir,
     automationPrefix: expectedFleet.automationPrefix,
   });
 
+  const runDisplays = await resolveFleetRunDisplays(projectRoot, expectedFleet);
+
   const expectedGroups = createAutomationGroupBins();
 
   const comparisons = compareAutomationFleet({
@@ -177,17 +189,23 @@ export async function inspectCodexAutomationFleet(input) {
         expected,
         comparison,
         now,
+        runDisplay: runDisplays.get(expected.id),
       })
     );
   }
 
   for (const unsupported of expectedFleet.unsupported) {
+    const runDisplay = runDisplays.get(unsupported.id);
     assignToAutomationGroup(expectedGroups, unsupported.group, {
       id: unsupported.automationId,
       status: "UNSUPPORTED",
       summary: unsupported.reason,
       expectedCadence: unsupported.expectedCadence,
       observed: "No automation is expected for this repo/runtime combination.",
+      runbook: runDisplay?.runbook,
+      lastOutcome: runDisplay?.lastOutcome,
+      outcomeHistory: runDisplay?.outcomeHistory,
+      olderRecordCount: runDisplay?.olderRecordCount,
     });
   }
 
@@ -374,15 +392,43 @@ async function readCodexAutomation(automationDir) {
   };
 }
 
+/**
+ * Read the read-only run-history display for every expected and unsupported
+ * loop up front, keyed by the loop's short id, so the per-entry item builder
+ * stays synchronous.
+ *
+ * @param {string} projectRoot
+ * @param {ExpectedFleet} expectedFleet
+ * @returns {Promise<Map<string, import("./automation-status-run-history.mjs").AutomationRunDisplay>>}
+ */
+async function resolveFleetRunDisplays(projectRoot, expectedFleet) {
+  const entries = [...expectedFleet.expected, ...expectedFleet.unsupported];
+  const displays = await Promise.all(
+    entries.map(entry =>
+      resolveAutomationRunDisplay({
+        projectRoot,
+        loopId: entry.id,
+        runbookPath: entry.runbookPath,
+      })
+    )
+  );
+  return new Map(entries.map((entry, index) => [entry.id, displays[index]]));
+}
+
 function createObservedStatusItem(input) {
   const expected = input.expected;
   const comparison = input.comparison;
   const observed = comparison.observedAutomation;
+  const runDisplay = input.runDisplay;
   const runSignal = classifyAutomationRunSignal({
     expected,
     observedAutomation: observed,
     now: input.now,
   });
+  // Three or more consecutive recorded recovery-required runs flip the loop to
+  // FAILING even when the scheduler entry looks fine — the fleet-health signal
+  // the raw backing store cannot see. It wins over the scheduler run-signal.
+  const escalation = resolveRecoveryEscalation(runDisplay);
 
   const observedDetails = [comparison.observed];
   if (observed?.status) {
@@ -401,20 +447,30 @@ function createObservedStatusItem(input) {
   }
 
   const status =
+    escalation?.status ??
     runSignal?.status ??
     /** @type {"HEALTHY" | "MISSING" | "DRIFTED"} */ (comparison.status);
 
   return {
     id: expected.automationId,
     status,
-    summary: composeAutomationSummary({
-      comparison,
-      runSignal,
-    }),
+    summary: escalation
+      ? escalation.summary
+      : composeAutomationSummary({
+          comparison,
+          runSignal,
+        }),
     expectedCadence: expected.expectedCadence,
     expectedCommand: expected.expectedCommand,
     observed: observedDetails.join(" "),
-    remediation: runSignal?.remediation ?? comparison.remediation,
+    runbook: runDisplay?.runbook,
+    lastOutcome: runDisplay?.lastOutcome,
+    outcomeHistory: runDisplay?.outcomeHistory,
+    olderRecordCount: runDisplay?.olderRecordCount,
+    remediation:
+      escalation?.remediation ??
+      runSignal?.remediation ??
+      comparison.remediation,
   };
 }
 

package/plugins/lisa-cursor/scripts/automation-status-report.mjs

@@ -26,12 +26,22 @@ export const AUTOMATION_FLEET_VERDICTS = [
  * @typedef {"HEALTHY" | "PARTIAL_SUPPORT" | "ATTENTION_NEEDED"} AutomationFleetVerdict
  *
  * @typedef {{
+ *   readonly ts: string
+ *   readonly outcome: string
+ *   readonly summary: string
+ * }} AutomationLastOutcome
+ *
+ * @typedef {{
  *   readonly id: string
  *   readonly status: AutomationHealthStatus
  *   readonly summary: string
  *   readonly expectedCadence?: string
  *   readonly expectedCommand?: string
  *   readonly observed?: string
+ *   readonly runbook?: string
+ *   readonly lastOutcome?: AutomationLastOutcome
+ *   readonly outcomeHistory?: readonly string[]
+ *   readonly olderRecordCount?: number
  *   readonly remediation?: string
  * }} AutomationStatusItem
  *
@@ -130,6 +140,26 @@ export function renderAutomationStatusReport(input) {
       if (item.observed) {
         lines.push(`  Observed: ${item.observed}`);
       }
+      // Contract + run-history block: observable facts about where the runbook
+      // lives and how recent runs ended, kept between Observed and Remediation
+      // so operators read the state before the fix. Every line degrades
+      // explicitly — an absent runbook or empty history is stated, never blank.
+      if (item.runbook !== undefined) {
+        lines.push(`  Runbook: ${item.runbook}`);
+        lines.push(
+          item.lastOutcome
+            ? `  Last run: ${item.lastOutcome.outcome} — ${item.lastOutcome.summary} (${item.lastOutcome.ts})`
+            : "  Last run: no recorded runs yet"
+        );
+        if (item.outcomeHistory && item.outcomeHistory.length > 0) {
+          const historyLine = `  History: ${item.outcomeHistory.join(", ")} (newest first)`;
+          lines.push(
+            item.olderRecordCount && item.olderRecordCount > 0
+              ? `${historyLine}; … and ${item.olderRecordCount} older records`
+              : historyLine
+          );
+        }
+      }
       if (item.remediation) {
         lines.push(`  Remediation: ${item.remediation}`);
       }

package/plugins/lisa-cursor/scripts/automation-status-run-history.mjs

@@ -0,0 +1,185 @@
+#!/usr/bin/env node
+/**
+ * Shared read-only run-history projection for `/lisa:automation-status`.
+ *
+ * The durable run-outcome substrate (RBC-3, `automation-run-record.mjs`) stores
+ * one bounded JSONL file per registered loop under `.lisa/automations/runs/`.
+ * This module turns that substrate — plus the checked-in per-loop runbook
+ * (`automation-runbook-contract`) — into the per-item display fields both
+ * runtime adapters attach to their status rows, so the Codex and Claude
+ * adapters never diverge on how the contract, last outcome, and bounded history
+ * are rendered.
+ *
+ * STRICTLY READ-ONLY: it stats the runbook and reads the run-record file. It
+ * never creates the runs directory, a runbook, or a run record — a status read
+ * must leave the fleet byte-identical.
+ */
+
+import fs from "node:fs/promises";
+import path from "node:path";
+
+import {
+  automationRunRecordPath,
+  readAutomationRunRecords,
+} from "./automation-run-record.mjs";
+
+/**
+ * Inline history cap: the newest N outcomes are listed, the remainder is only
+ * counted. A bounded view of an already-bounded file, with the trim stated in
+ * the output rather than applied silently.
+ */
+export const AUTOMATION_RUN_HISTORY_DISPLAY_LIMIT = 5;
+
+/** The recorded outcome that signals a loop could not finish its own work. */
+export const RECOVERY_REQUIRED_OUTCOME = "recovery-required";
+
+/**
+ * How many trailing `recovery-required` runs in a row flip a loop to `FAILING`.
+ * The RBC-5 acceptance criteria pin "last three runs" as the escalation
+ * trigger; that wins over the ticket prose's looser "five in a row" pattern.
+ */
+export const RECOVERY_REQUIRED_ESCALATION_THRESHOLD = 3;
+
+const RUNBOOK_NOT_SCAFFOLDED_LINE =
+  "not scaffolded — run /lisa:setup-automations";
+const NO_RECORDED_RUNS_LINE = "no recorded runs yet";
+
+/**
+ * @typedef {{
+ *   readonly ts: string
+ *   readonly outcome: string
+ *   readonly summary: string
+ * }} AutomationLastOutcome
+ *
+ * @typedef {{
+ *   readonly runbook: string
+ *   readonly lastOutcome?: AutomationLastOutcome
+ *   readonly outcomeHistory: readonly string[]
+ *   readonly olderRecordCount: number
+ *   readonly recoveryRequiredStreak: number
+ *   readonly recoverySummaries: readonly string[]
+ *   readonly skippedCorruptLines: number
+ * }} AutomationRunDisplay
+ */
+
+/**
+ * Resolve the read-only run-history display fields for one registered loop.
+ *
+ * @param {{
+ *   readonly projectRoot?: string
+ *   readonly loopId: string
+ *   readonly runbookPath?: string
+ * }} input
+ * @returns {Promise<AutomationRunDisplay>}
+ */
+export async function resolveAutomationRunDisplay(input) {
+  const projectRoot = path.resolve(input.projectRoot ?? process.cwd());
+  const runbook = await resolveRunbookLine(projectRoot, input.runbookPath);
+  const { records, skippedCorruptLines } = await readAutomationRunRecords(
+    automationRunRecordPath(projectRoot, input.loopId)
+  );
+
+  const newestFirst = records.toReversed();
+  const latest = records.at(-1);
+  const lastOutcome = latest
+    ? { ts: latest.ts, outcome: latest.outcome, summary: latest.summary }
+    : undefined;
+  const outcomeHistory = newestFirst
+    .slice(0, AUTOMATION_RUN_HISTORY_DISPLAY_LIMIT)
+    .map(record => record.outcome);
+  const olderRecordCount = Math.max(
+    0,
+    records.length - AUTOMATION_RUN_HISTORY_DISPLAY_LIMIT
+  );
+  const trailingRecovery = collectTrailingRecoveryRuns(records);
+
+  return {
+    runbook,
+    lastOutcome,
+    outcomeHistory,
+    olderRecordCount,
+    recoveryRequiredStreak: trailingRecovery.length,
+    recoverySummaries: trailingRecovery.map(record => record.summary),
+    skippedCorruptLines,
+  };
+}
+
+/**
+ * Force a `FAILING` verdict when a loop has recorded three or more consecutive
+ * `recovery-required` runs — the fleet-health signal the scheduler surface
+ * cannot see on its own. Returns a run-signal-shaped escalation, or null when
+ * the streak is below the threshold.
+ *
+ * @param {AutomationRunDisplay | undefined} runDisplay
+ * @returns {{ readonly status: "FAILING", readonly summary: string, readonly remediation: string } | null}
+ */
+export function resolveRecoveryEscalation(runDisplay) {
+  if (
+    !runDisplay ||
+    runDisplay.recoveryRequiredStreak < RECOVERY_REQUIRED_ESCALATION_THRESHOLD
+  ) {
+    return null;
+  }
+
+  // Number the citations rather than joining on a delimiter: recorded summaries
+  // carry their own `;` and `.`, so a `; `-separated list is ambiguous. `(1) …
+  // (2) …` stays legible no matter what punctuation a summary contains.
+  const citations = runDisplay.recoverySummaries
+    .map((summary, index) => `(${index + 1}) ${summary}`)
+    .join(" ");
+  return {
+    status: "FAILING",
+    summary: `the last ${runDisplay.recoveryRequiredStreak} recorded runs all required recovery`,
+    remediation: `Repeated recovery-required runs — ${citations} Inspect the loop's runbook and the cited runs, fix the recurring failure, then let the next scheduled run confirm recovery.`,
+  };
+}
+
+/**
+ * @param {string} projectRoot
+ * @param {string | undefined} runbookPath
+ * @returns {Promise<string>}
+ */
+async function resolveRunbookLine(projectRoot, runbookPath) {
+  if (!runbookPath) {
+    return RUNBOOK_NOT_SCAFFOLDED_LINE;
+  }
+  const scaffolded = await runbookIsScaffolded(
+    path.join(projectRoot, runbookPath)
+  );
+  return scaffolded ? runbookPath : RUNBOOK_NOT_SCAFFOLDED_LINE;
+}
+
+/**
+ * Read-only existence check for a checked-in runbook. Any stat failure — the
+ * file is absent, unreadable, or not a regular file — degrades to "not
+ * scaffolded" so the status read never implies health it could not confirm and
+ * never throws mid-report.
+ *
+ * @param {string} runbookAbsolutePath
+ * @returns {Promise<boolean>}
+ */
+async function runbookIsScaffolded(runbookAbsolutePath) {
+  try {
+    const stat = await fs.stat(runbookAbsolutePath);
+    return stat.isFile();
+  } catch {
+    return false;
+  }
+}
+
+/**
+ * @param {readonly import("./automation-run-record.mjs").AutomationRunRecord[]} records
+ * @returns {readonly import("./automation-run-record.mjs").AutomationRunRecord[]} trailing recovery-required runs, newest first
+ */
+function collectTrailingRecoveryRuns(records) {
+  const trailing = [];
+  for (let index = records.length - 1; index >= 0; index -= 1) {
+    if (records[index].outcome !== RECOVERY_REQUIRED_OUTCOME) {
+      break;
+    }
+    trailing.push(records[index]);
+  }
+  return trailing;
+}
+
+export { RUNBOOK_NOT_SCAFFOLDED_LINE, NO_RECORDED_RUNS_LINE };