diff --git a/actions/setup/js/evaluate_outcomes.cjs b/actions/setup/js/evaluate_outcomes.cjs
index 52be0fc3089..a98a8501348 100644
--- a/actions/setup/js/evaluate_outcomes.cjs
+++ b/actions/setup/js/evaluate_outcomes.cjs
@@ -455,7 +455,7 @@ function evaluateAddComment(item, itemRepo, timestamp, out, apiGet, nowMs) {
  * @returns {EvalResult}
  */
 function evaluateAddLabels(item, itemRepo, timestamp, out, apiGet, nowMs) {
-  const num = parseIssueNumberFromURL(item.url || "");
+  const num = getItemNumber(item);
   if (!num) {
     out.result = "unknown";
     out.detail = "unknown: issue number not found";
@@ -463,8 +463,9 @@ function evaluateAddLabels(item, itemRepo, timestamp, out, apiGet, nowMs) {
     return out;
   }
 
-  const hasLabelsBefore = Object.hasOwn(item, "labelsBefore") && Array.isArray(item.labelsBefore);
-  const labelsBefore = hasLabelsBefore ? normalizeLabels(item.labelsBefore) : [];
+  const persistedBefore = item.before_state?.labels ?? item.labelsBefore;
+  const hasLabelsBefore = Array.isArray(persistedBefore);
+  const labelsBefore = hasLabelsBefore ? normalizeLabels(persistedBefore) : [];
   const labelsAdded = normalizeLabels(item.labelsAdded);
   const fallbackLabels = normalizeLabels(item.labels);
   const effectiveLabelsAdded = labelsAdded.length > 0 ? labelsAdded : fallbackLabels;
@@ -534,6 +535,170 @@ function evaluateAddLabels(item, itemRepo, timestamp, out, apiGet, nowMs) {
   return out;
 }
 
+/**
+ * Evaluate `close_issue`.
+ * @param {object} item
+ * @param {string} defaultRepo
+ * @param {(endpoint: string) => any} api
+ * @param {number} nowMs
+ * @returns {EvalResult}
+ */
+function evaluateCloseIssue(item, defaultRepo, api = ghAPI, nowMs = Date.now()) {
+  const repo = getItemRepo(item, defaultRepo);
+  const number = getItemNumber(item);
+  const timestamp = item.timestamp || "";
+  /** @type {EvalResult} */
+  const out = {
+    result: "unknown",
+    outcome_status: "unknown",
+    evidence_strength: "weak",
+    signal: "unknown",
+    detail: "",
+    resolution_sec: null,
+    pending_age_sec: null,
+    review_comments: null,
+    changed_files: null,
+    additions: null,
+    deletions: null,
+    reactions_total: null,
+    reactions_positive: null,
+    reactions_negative: null,
+    comments: null,
+    zero_touch: false,
+  };
+
+  if (!repo || !number) {
+    out.detail = "missing issue reference";
+    return out;
+  }
+
+  const issue = api(`repos/${repo}/issues/${number}`);
+  if (!issue || typeof issue.state !== "string") {
+    out.detail = "api error";
+    setPendingAge(out, timestamp, nowMs);
+    return out;
+  }
+
+  out.comments = typeof issue.comments === "number" ? issue.comments : null;
+  if (issue.reactions && typeof issue.reactions === "object") {
+    const summary = summarizeReactions(issue.reactions);
+    out.reactions_total = summary.total;
+    out.reactions_positive = summary.positive;
+    out.reactions_negative = summary.negative;
+  }
+
+  if (issue.state === "closed") {
+    out.result = "accepted";
+    out.outcome_status = "accepted";
+    out.evidence_strength = "strong";
+    out.signal = "closed";
+    out.detail = "closed";
+    if (issue.created_at && issue.closed_at) {
+      out.resolution_sec = secondsBetween(issue.created_at, issue.closed_at);
+    }
+    return out;
+  }
+
+  out.result = "rejected";
+  out.outcome_status = "rejected";
+  out.evidence_strength = "strong";
+  out.signal = "not_closed";
+  out.detail = "not_closed";
+  return out;
+}
+
+/**
+ * Evaluate `close_pull_request`.
+ * @param {object} item
+ * @param {string} defaultRepo
+ * @param {(endpoint: string) => any} api
+ * @param {number} nowMs
+ * @returns {EvalResult}
+ */
+function evaluateClosePullRequest(item, defaultRepo, api = ghAPI, nowMs = Date.now()) {
+  const repo = getItemRepo(item, defaultRepo);
+  const number = getItemNumber(item);
+  const timestamp = item.timestamp || "";
+  /** @type {EvalResult} */
+  const out = {
+    result: "unknown",
+    outcome_status: "unknown",
+    evidence_strength: "weak",
+    signal: "unknown",
+    detail: "",
+    resolution_sec: null,
+    pending_age_sec: null,
+    review_comments: null,
+    changed_files: null,
+    additions: null,
+    deletions: null,
+    reactions_total: null,
+    reactions_positive: null,
+    reactions_negative: null,
+    comments: null,
+    zero_touch: false,
+  };
+
+  if (!repo || !number) {
+    out.detail = "missing pull request reference";
+    return out;
+  }
+
+  const pullRequest = api(`repos/${repo}/pulls/${number}`);
+  if (!pullRequest || typeof pullRequest.state !== "string") {
+    out.detail = "api error";
+    setPendingAge(out, timestamp, nowMs);
+    return out;
+  }
+
+  out.review_comments = typeof pullRequest.review_comments === "number" ? pullRequest.review_comments : null;
+  out.changed_files = typeof pullRequest.changed_files === "number" ? pullRequest.changed_files : null;
+  out.additions = typeof pullRequest.additions === "number" ? pullRequest.additions : null;
+  out.deletions = typeof pullRequest.deletions === "number" ? pullRequest.deletions : null;
+  out.comments = typeof pullRequest.comments === "number" ? pullRequest.comments : null;
+  if (pullRequest.reactions && typeof pullRequest.reactions === "object") {
+    const summary = summarizeReactions(pullRequest.reactions);
+    out.reactions_total = summary.total;
+    out.reactions_positive = summary.positive;
+    out.reactions_negative = summary.negative;
+  }
+
+  // A merged PR is rejected because close_pull_request verifies that the PR
+  // remained closed without being merged. Merging is a different terminal
+  // state than closing, so it invalidates the close outcome.
+  if (pullRequest.merged === true || pullRequest.merged_at != null) {
+    out.result = "rejected";
+    out.outcome_status = "rejected";
+    out.evidence_strength = "strong";
+    out.signal = "closed_by_merge";
+    out.detail = "merged";
+    if (pullRequest.created_at && pullRequest.merged_at) {
+      out.resolution_sec = secondsBetween(pullRequest.created_at, pullRequest.merged_at);
+    }
+    return out;
+  }
+  // Accepted means the PR is closed and unmerged, which is the durable
+  // terminal state that close_pull_request validates.
+  if (pullRequest.state === "closed") {
+    out.result = "accepted";
+    out.outcome_status = "accepted";
+    out.evidence_strength = "strong";
+    out.signal = "closed";
+    out.detail = "closed";
+    if (pullRequest.created_at && pullRequest.closed_at) {
+      out.resolution_sec = secondsBetween(pullRequest.created_at, pullRequest.closed_at);
+    }
+    return out;
+  }
+
+  out.result = "rejected";
+  out.outcome_status = "rejected";
+  out.evidence_strength = "strong";
+  out.signal = "not_closed";
+  out.detail = "not_closed";
+  return out;
+}
+
 /**
  * Normalize legacy result/detail pairs into the shared outcome model.
  * @param {string} result
@@ -596,9 +761,18 @@ function normalizeOutcome(result, detail) {
   if (result === "accepted" && normalizedDetail.startsWith("merged")) {
     return { outcome_status: "accepted", evidence_strength: "strong", signal: "merged" };
   }
+  if (result === "accepted" && normalizedDetail === "closed") {
+    return { outcome_status: "accepted", evidence_strength: "strong", signal: "closed" };
+  }
   if (result === "rejected" && normalizedDetail === "closed") {
     return { outcome_status: "rejected", evidence_strength: "strong", signal: "closed" };
   }
+  if (result === "rejected" && normalizedDetail === "merged") {
+    return { outcome_status: "rejected", evidence_strength: "strong", signal: "closed_by_merge" };
+  }
+  if (result === "rejected" && normalizedDetail === "not_closed") {
+    return { outcome_status: "rejected", evidence_strength: "strong", signal: "not_closed" };
+  }
   if (result === "pending" && normalizedDetail === "open") {
     return { outcome_status: "pending", evidence_strength: "medium", signal: "open" };
   }
@@ -1187,6 +1361,21 @@ function evaluateItem(item, defaultRepo, apiOrOptions) {
   if (type === "update_pull_request") {
     return evaluateUpdatePullRequest(item, defaultRepo, ghAPIFn);
   }
+  if (type === "add_labels") {
+    return evaluateAddLabels(item, itemRepo, timestamp, out, ghAPIFn, nowMs);
+  }
+  if (type === "close_issue") {
+    return evaluateCloseIssue(item, defaultRepo, ghAPIFn, nowMs);
+  }
+  if (type === "close_pull_request") {
+    return evaluateClosePullRequest(item, defaultRepo, ghAPIFn, nowMs);
+  }
+  if (type === "create_issue") {
+    return evaluateCreateIssue(item, itemRepo, timestamp, out, ghAPIFn, nowMs);
+  }
+  if (type === "add_comment") {
+    return evaluateAddComment(item, itemRepo, timestamp, out, ghAPIFn, nowMs);
+  }
 
   if (!url) {
     if (item.type === "add_reviewer") {
@@ -1206,15 +1395,6 @@ function evaluateItem(item, defaultRepo, apiOrOptions) {
   if (type === "submit_pull_request_review") {
     return evaluateSubmitPullRequestReview(item, defaultRepo, ghAPIFn);
   }
-  if (type === "create_issue") {
-    return evaluateCreateIssue(item, itemRepo, timestamp, out, ghAPIFn, nowMs);
-  }
-  if (type === "add_comment") {
-    return evaluateAddComment(item, itemRepo, timestamp, out, ghAPIFn, nowMs);
-  }
-  if (type === "add_labels") {
-    return evaluateAddLabels(item, itemRepo, timestamp, out, ghAPIFn, nowMs);
-  }
 
   // Issues / issue-comments
   const issueMatch = url.match(/\/(?:issues|pull)\/(\d+)/);
@@ -1944,6 +2124,8 @@ module.exports = {
   evaluateCreateIssue,
   evaluateAddComment,
   evaluateAddLabels,
+  evaluateCloseIssue,
+  evaluateClosePullRequest,
   evaluateCreatePullRequestOutcome,
   evaluatePushToPullRequestBranchOutcome,
   normalizeOutcome,
diff --git a/actions/setup/js/evaluate_outcomes.test.cjs b/actions/setup/js/evaluate_outcomes.test.cjs
index 72487ce817a..010f1b95bf1 100644
--- a/actions/setup/js/evaluate_outcomes.test.cjs
+++ b/actions/setup/js/evaluate_outcomes.test.cjs
@@ -142,7 +142,8 @@ describe("evaluate_outcomes type-specific evaluators", () => {
   it("evaluates add_labels retention using persisted before-state labels", () => {
     const item = {
       type: "add_labels",
-      url: "https://github.com/acme/repo/issues/42",
+      number: 42,
+      repo: "acme/repo",
       timestamp: "2026-05-25T00:00:00Z",
       labelsBefore: ["bug"],
       labelsAdded: ["bug", "triage"],
@@ -164,6 +165,192 @@ describe("evaluate_outcomes type-specific evaluators", () => {
     expect(evaluateItem({ ...item, labelsBefore: [] }, "acme/repo", { ghAPI: retained, nowMs: Date.parse("2026-05-27T00:00:00Z") }).detail).toBe("accepted:strong");
   });
 
+  it("evaluates add_labels retention using production before_state.labels manifest shape", () => {
+    // Production manifest from add_labels.cjs stores labels in before_state.labels, not labelsBefore.
+    const item = {
+      type: "add_labels",
+      number: 42,
+      repo: "acme/repo",
+      timestamp: "2026-05-25T00:00:00Z",
+      before_state: { labels: ["bug"] },
+      labelsAdded: ["bug", "triage"],
+    };
+
+    const retained = createAPIStub({
+      "repos/acme/repo/issues/42/labels": [{ name: "bug" }, { name: "triage" }],
+    });
+    expect(evaluateItem(item, "acme/repo", { ghAPI: retained, nowMs: Date.parse("2026-05-27T00:00:00Z") }).detail).toBe("accepted:strong");
+  });
+
+  it("returns unknown for add_labels when the item number is missing", () => {
+    const missingNumber = evaluateItem(
+      {
+        type: "add_labels",
+        repo: "acme/repo",
+        timestamp: "2026-05-25T00:00:00Z",
+        labelsBefore: ["bug"],
+        labelsAdded: ["bug", "triage"],
+      },
+      "acme/repo",
+      {
+        ghAPI: createAPIStub({
+          "repos/acme/repo/issues/42/labels": [{ name: "bug" }, { name: "triage" }],
+        }),
+        nowMs: Date.parse("2026-05-27T00:00:00Z"),
+      }
+    );
+    expect(missingNumber.detail).toBe("unknown: issue number not found");
+  });
+
+  it("evaluates close_issue using persisted repo/number metadata", () => {
+    const item = {
+      type: "close_issue",
+      number: 77,
+      repo: "acme/repo",
+      timestamp: "2026-05-26T00:00:00Z",
+    };
+
+    const closed = createAPIStub({
+      "repos/acme/repo/issues/77": {
+        state: "closed",
+        comments: 1,
+        reactions: { total_count: 2 },
+        created_at: "2026-05-26T00:00:00Z",
+        closed_at: "2026-05-26T00:10:00Z",
+      },
+    });
+    expect(evaluateItem(item, "acme/repo", { ghAPI: closed })).toMatchObject({
+      result: "accepted",
+      outcome_status: "accepted",
+      evidence_strength: "strong",
+      signal: "closed",
+      detail: "closed",
+    });
+
+    const open = createAPIStub({
+      "repos/acme/repo/issues/77": { state: "open", comments: 1, reactions: { total_count: 0 } },
+    });
+    expect(evaluateItem(item, "acme/repo", { ghAPI: open })).toMatchObject({
+      result: "rejected",
+      outcome_status: "rejected",
+      evidence_strength: "strong",
+      signal: "not_closed",
+      detail: "not_closed",
+    });
+  });
+
+  it("returns unknown for close_issue when the item reference is missing", () => {
+    const missingRef = evaluateItem({ type: "close_issue", timestamp: "2026-05-26T00:00:00Z" }, "acme/repo", { ghAPI: createAPIStub({}), nowMs: Date.parse("2026-05-27T00:00:00Z") });
+    expect(missingRef.result).toBe("unknown");
+    expect(missingRef.detail).toBe("missing issue reference");
+  });
+
+  it("returns unknown for close_issue on API error and sets pending_age_sec", () => {
+    const nowMs = Date.parse("2026-05-27T00:00:00Z");
+    const apiError = evaluateItem({ type: "close_issue", number: 77, repo: "acme/repo", timestamp: "2026-05-26T00:00:00Z" }, "acme/repo", { ghAPI: createAPIStub({}), nowMs });
+    expect(apiError.result).toBe("unknown");
+    expect(apiError.detail).toBe("api error");
+    expect(typeof apiError.pending_age_sec).toBe("number");
+  });
+
+  it("evaluates close_pull_request using persisted repo/number metadata", () => {
+    const item = {
+      type: "close_pull_request",
+      number: 55,
+      repo: "acme/repo",
+      timestamp: "2026-05-26T00:00:00Z",
+    };
+
+    const closed = createAPIStub({
+      "repos/acme/repo/pulls/55": {
+        state: "closed",
+        merged: false,
+        review_comments: 0,
+        changed_files: 1,
+        additions: 3,
+        deletions: 1,
+        comments: 0,
+        created_at: "2026-05-26T00:00:00Z",
+        closed_at: "2026-05-26T00:10:00Z",
+      },
+    });
+    expect(evaluateItem(item, "acme/repo", { ghAPI: closed })).toMatchObject({
+      result: "accepted",
+      outcome_status: "accepted",
+      evidence_strength: "strong",
+      signal: "closed",
+      detail: "closed",
+    });
+
+    const merged = createAPIStub({
+      "repos/acme/repo/pulls/55": {
+        state: "closed",
+        merged: true,
+        review_comments: 0,
+        changed_files: 1,
+        additions: 3,
+        deletions: 1,
+        comments: 0,
+        created_at: "2026-05-26T00:00:00Z",
+        merged_at: "2026-05-26T00:10:00Z",
+      },
+    });
+    expect(evaluateItem(item, "acme/repo", { ghAPI: merged })).toMatchObject({
+      result: "rejected",
+      outcome_status: "rejected",
+      evidence_strength: "strong",
+      signal: "closed_by_merge",
+      detail: "merged",
+    });
+
+    const open = createAPIStub({
+      "repos/acme/repo/pulls/55": { state: "open", merged: false, review_comments: 0 },
+    });
+    expect(evaluateItem(item, "acme/repo", { ghAPI: open })).toMatchObject({
+      result: "rejected",
+      outcome_status: "rejected",
+      evidence_strength: "strong",
+      signal: "not_closed",
+      detail: "not_closed",
+    });
+  });
+
+  it("detects merged PR via merged_at when merged flag is absent", () => {
+    const item = {
+      type: "close_pull_request",
+      number: 55,
+      repo: "acme/repo",
+      timestamp: "2026-05-26T00:00:00Z",
+    };
+    const mergedAtOnly = createAPIStub({
+      "repos/acme/repo/pulls/55": {
+        state: "closed",
+        merged: false,
+        merged_at: "2026-05-26T00:10:00Z",
+        created_at: "2026-05-26T00:00:00Z",
+      },
+    });
+    expect(evaluateItem(item, "acme/repo", { ghAPI: mergedAtOnly })).toMatchObject({
+      result: "rejected",
+      signal: "closed_by_merge",
+      detail: "merged",
+    });
+  });
+
+  it("returns unknown for close_pull_request when the item reference is missing", () => {
+    const missingRef = evaluateItem({ type: "close_pull_request", timestamp: "2026-05-26T00:00:00Z" }, "acme/repo", { ghAPI: createAPIStub({}), nowMs: Date.parse("2026-05-27T00:00:00Z") });
+    expect(missingRef.result).toBe("unknown");
+    expect(missingRef.detail).toBe("missing pull request reference");
+  });
+
+  it("returns unknown for close_pull_request on API error and sets pending_age_sec", () => {
+    const nowMs = Date.parse("2026-05-27T00:00:00Z");
+    const apiError = evaluateItem({ type: "close_pull_request", number: 55, repo: "acme/repo", timestamp: "2026-05-26T00:00:00Z" }, "acme/repo", { ghAPI: createAPIStub({}), nowMs });
+    expect(apiError.result).toBe("unknown");
+    expect(apiError.detail).toBe("api error");
+    expect(typeof apiError.pending_age_sec).toBe("number");
+  });
+
   it("evaluates update_issue retained and reverted states from persisted execution metadata", () => {
     const retained = evaluateItem(
       {
diff --git a/pkg/cli/outcome_eval.go b/pkg/cli/outcome_eval.go
index fcd9908aacd..d2af77bcd92 100644
--- a/pkg/cli/outcome_eval.go
+++ b/pkg/cli/outcome_eval.go
@@ -92,6 +92,7 @@ var outcomeEvaluators = map[string]outcomeEvaluator{
 	"update_pull_request":                   evalUpdatePullRequest,
 	"add_comment":                           evalAddComment,
 	"add_labels":                            evalAddLabels,
+	"replace_label":                         evalReplaceLabel,
 	"assign_to_agent":                       evalAssignToAgent,
 	"close_issue":                           evalCloseSticky,
 	"close_pull_request":                    evalCloseSticky,
diff --git a/pkg/cli/outcome_eval_label.go b/pkg/cli/outcome_eval_label.go
index 21be3ab8a0b..0dd4fe65348 100644
--- a/pkg/cli/outcome_eval_label.go
+++ b/pkg/cli/outcome_eval_label.go
@@ -2,12 +2,139 @@ package cli
 
 import (
 	"fmt"
+	"slices"
 
 	"github.com/github/gh-aw/pkg/logger"
 )
 
 var outcomeEvalLabelLog = logger.New("cli:outcome_eval_label")
 
+// evalReplaceLabel checks whether the label delta applied at execution time is
+// still intact in the current issue state.  It computes the set of labels added
+// and removed by the replacement and verifies only that delta, ignoring any
+// unrelated labels that may have been added or removed since execution.
+func evalReplaceLabel(item CreatedItemReport, repoOverride string) OutcomeReport {
+	repo := resolveItemRepo(item, repoOverride)
+	num := resolveItemNumber(item)
+	outcomeEvalLabelLog.Printf("Evaluating replace_label outcome: repo=%s, number=%d", repo, num)
+	report := OutcomeReport{
+		Type:         item.Type,
+		ObjectURL:    item.URL,
+		ObjectNumber: num,
+		Repo:         repo,
+	}
+	if num == 0 || repo == "" || item.BeforeState == nil || item.AfterState == nil {
+		report.Result = OutcomeUnknown
+		report.Detail = "missing execution state"
+		report.OutcomeEvaluation = OutcomeEvaluation{
+			OutcomeStatus:    OutcomeStatusUnknown,
+			EvidenceStrength: EvidenceNone,
+			Signal:           "missing_execution_state",
+		}
+		return report
+	}
+
+	beforeLabels := mutableStringSlice(item.BeforeState["labels"])
+	afterLabels := mutableStringSlice(item.AfterState["labels"])
+
+	// Compute the replacement delta: labels added and labels removed.
+	added := labelSetDiff(afterLabels, beforeLabels)
+	removed := labelSetDiff(beforeLabels, afterLabels)
+
+	if len(added) == 0 && len(removed) == 0 {
+		report.Result = OutcomeUnknown
+		report.Detail = "no label delta"
+		report.OutcomeEvaluation = OutcomeEvaluation{
+			OutcomeStatus:    OutcomeStatusUnknown,
+			EvidenceStrength: EvidenceNone,
+			Signal:           "no_state_delta",
+		}
+		return report
+	}
+
+	currentState, _, err := extractCurrentIssueUpdateState(repo, num)
+	if err != nil {
+		report.Result = OutcomeError
+		report.EvalError = err.Error()
+		return report
+	}
+	currentLabels := mutableStringSlice(currentState["labels"])
+
+	// The replacement is retained when all added labels are still present and
+	// all removed labels are still absent, regardless of any other label changes.
+	addedRetained := labelSetContainsAll(currentLabels, added)
+	removedStillAbsent := !labelSetContainsAny(currentLabels, removed)
+
+	if addedRetained && removedStillAbsent {
+		report.Result = OutcomeAccepted
+		report.Detail = "label replacement retained"
+		report.OutcomeEvaluation = OutcomeEvaluation{
+			OutcomeStatus:    OutcomeStatusAccepted,
+			EvidenceStrength: EvidenceMedium,
+			Signal:           "state_retained",
+		}
+		return report
+	}
+
+	// Reverted: all added labels are gone and all removed labels are back.
+	addedReverted := !labelSetContainsAny(currentLabels, added)
+	removedBack := labelSetContainsAll(currentLabels, removed)
+	if addedReverted && removedBack {
+		report.Result = OutcomeRejected
+		report.Detail = "label replacement reverted"
+		report.OutcomeEvaluation = OutcomeEvaluation{
+			OutcomeStatus:    OutcomeStatusRejected,
+			EvidenceStrength: EvidenceStrong,
+			Signal:           "state_reverted",
+		}
+		return report
+	}
+
+	report.Result = OutcomeRejected
+	report.Detail = "label replacement replaced"
+	report.OutcomeEvaluation = OutcomeEvaluation{
+		OutcomeStatus:    OutcomeStatusRejected,
+		EvidenceStrength: EvidenceStrong,
+		Signal:           "state_replaced",
+	}
+	return report
+}
+
+// labelSetDiff returns the elements of a that are not in b.
+// Both slices must be sorted (as produced by mutableStringSlice).
+// Uses binary search for O(n log m) performance.
+func labelSetDiff(a, b []string) []string {
+	var out []string
+	for _, v := range a {
+		if _, found := slices.BinarySearch(b, v); !found {
+			out = append(out, v)
+		}
+	}
+	return out
+}
+
+// labelSetContainsAll reports whether current contains every element of want.
+// Both slices must be sorted. Uses binary search for O(n log m) performance.
+func labelSetContainsAll(current, want []string) bool {
+	for _, v := range want {
+		if _, found := slices.BinarySearch(current, v); !found {
+			return false
+		}
+	}
+	return true
+}
+
+// labelSetContainsAny reports whether current contains at least one element of want.
+// Both slices must be sorted. Uses binary search for O(n log m) performance.
+func labelSetContainsAny(current, want []string) bool {
+	for _, v := range want {
+		if _, found := slices.BinarySearch(current, v); found {
+			return true
+		}
+	}
+	return false
+}
+
 // evalAddLabels checks whether labels added by the workflow are still present.
 func evalAddLabels(item CreatedItemReport, repoOverride string) OutcomeReport {
 	repo := resolveItemRepo(item, repoOverride)
diff --git a/pkg/cli/outcome_eval_update.go b/pkg/cli/outcome_eval_update.go
index d4fcf559e50..4a888e06d7b 100644
--- a/pkg/cli/outcome_eval_update.go
+++ b/pkg/cli/outcome_eval_update.go
@@ -155,6 +155,8 @@ func mutableTrackedFields(itemType string) []string {
 		return []string{"title", "body_hash", "state", "labels", "assignees"}
 	case "update_pull_request":
 		return []string{"title", "body_hash", "state", "base", "draft", "head_sha"}
+	case "replace_label":
+		return []string{"labels"}
 	default:
 		return nil
 	}
diff --git a/pkg/cli/outcome_eval_update_test.go b/pkg/cli/outcome_eval_update_test.go
index fb80f225b9a..914854ea18d 100644
--- a/pkg/cli/outcome_eval_update_test.go
+++ b/pkg/cli/outcome_eval_update_test.go
@@ -191,3 +191,105 @@ func TestEvalRetainedUpdateMissingExecutionStateUsesEvidenceNone(t *testing.T) {
 	assert.Equal(t, EvidenceNone, report.EvidenceStrength)
 	assert.Equal(t, "missing_execution_state", report.Signal)
 }
+
+func TestEvalReplaceLabelRetained(t *testing.T) {
+	old := outcomeUpdateGHAPIGet
+	t.Cleanup(func() {
+		outcomeUpdateGHAPIGet = old
+	})
+	outcomeUpdateGHAPIGet = func(endpoint string, repo string) (map[string]any, error) {
+		return map[string]any{
+			"labels": []any{
+				map[string]any{"name": "triage"},
+				map[string]any{"name": "done"},
+			},
+		}, nil
+	}
+
+	report := evalReplaceLabel(CreatedItemReport{
+		Type:   "replace_label",
+		Number: 12,
+		Repo:   "owner/repo",
+		BeforeState: map[string]any{
+			"labels": []any{"triage", "in-progress"},
+		},
+		AfterState: map[string]any{
+			"labels": []any{"triage", "done"},
+		},
+	}, "owner/repo")
+
+	assert.Equal(t, OutcomeAccepted, report.Result)
+	assert.Equal(t, OutcomeStatusAccepted, report.OutcomeStatus)
+	assert.Equal(t, EvidenceMedium, report.EvidenceStrength)
+	assert.Equal(t, "state_retained", report.Signal)
+}
+
+func TestEvalReplaceLabelRetainedWithExtraLabel(t *testing.T) {
+	// An unrelated label added after execution must not cause state_replaced.
+	// Before: [in-progress], After: [done], Current: [done, security]
+	// Delta: added=[done], removed=[in-progress]
+	// "done" is still present, "in-progress" still absent → accepted.
+	old := outcomeUpdateGHAPIGet
+	t.Cleanup(func() {
+		outcomeUpdateGHAPIGet = old
+	})
+	outcomeUpdateGHAPIGet = func(endpoint string, repo string) (map[string]any, error) {
+		return map[string]any{
+			"labels": []any{
+				map[string]any{"name": "done"},
+				map[string]any{"name": "security"},
+			},
+		}, nil
+	}
+
+	report := evalReplaceLabel(CreatedItemReport{
+		Type:   "replace_label",
+		Number: 12,
+		Repo:   "owner/repo",
+		BeforeState: map[string]any{
+			"labels": []any{"in-progress"},
+		},
+		AfterState: map[string]any{
+			"labels": []any{"done"},
+		},
+	}, "owner/repo")
+
+	assert.Equal(t, OutcomeAccepted, report.Result)
+	assert.Equal(t, OutcomeStatusAccepted, report.OutcomeStatus)
+	assert.Equal(t, EvidenceMedium, report.EvidenceStrength)
+	assert.Equal(t, "state_retained", report.Signal)
+}
+
+func TestEvalReplaceLabelReverted(t *testing.T) {
+	// Before: [in-progress], After: [done], Current: [in-progress]
+	// Delta: added=[done], removed=[in-progress]
+	// "done" absent, "in-progress" back → reverted.
+	old := outcomeUpdateGHAPIGet
+	t.Cleanup(func() {
+		outcomeUpdateGHAPIGet = old
+	})
+	outcomeUpdateGHAPIGet = func(endpoint string, repo string) (map[string]any, error) {
+		return map[string]any{
+			"labels": []any{
+				map[string]any{"name": "in-progress"},
+			},
+		}, nil
+	}
+
+	report := evalReplaceLabel(CreatedItemReport{
+		Type:   "replace_label",
+		Number: 12,
+		Repo:   "owner/repo",
+		BeforeState: map[string]any{
+			"labels": []any{"in-progress"},
+		},
+		AfterState: map[string]any{
+			"labels": []any{"done"},
+		},
+	}, "owner/repo")
+
+	assert.Equal(t, OutcomeRejected, report.Result)
+	assert.Equal(t, OutcomeStatusRejected, report.OutcomeStatus)
+	assert.Equal(t, EvidenceStrong, report.EvidenceStrength)
+	assert.Equal(t, "state_reverted", report.Signal)
+}
diff --git a/specs/intent-attribution-agent-governance.md b/specs/intent-attribution-agent-governance.md
index 5184edce5df..187d77399c6 100644
--- a/specs/intent-attribution-agent-governance.md
+++ b/specs/intent-attribution-agent-governance.md
@@ -92,6 +92,33 @@ These capabilities remain supported.
 
 In this specification, they are treated as an early implementation of **intent attribution**, not as proof of business impact.
 
+### Safeguards
+
+When governance policy resolution fails because `.github/objective-mapping.json`,
+`.github/intent-policy.json`, or equivalent policy inputs are missing, malformed,
+or produce no deterministic match, the implementation MUST fail closed to the
+safest policy defined in **Fail-Closed Behavior** above.
+
+An implementation MUST NOT silently reuse a stale cached policy decision when the
+current repository policy inputs cannot be resolved. The failure reason and the
+fallback to safe policy decision SHOULD be recorded in execution provenance.
+
+### Sync Notes
+
+Until repositories migrate fully to `.github/intent-policy.json`, `.github/objective-mapping.json`
+remains the authoritative label-to-intent source for attribution fallback and drift detection.
+This specification defines the precedence and sync expectations for that migration, but does
+not assign a mandatory repository-by-repository completion deadline.
+For this document, "fully migrate" means all active intent keys and governance rules are
+defined in `.github/intent-policy.json`, and runtime authorization no longer depends on
+reading `.github/objective-mapping.json` except for explicit backward-compatibility paths.
+
+Implementations SHOULD detect drift by comparing the active repository governance inputs —
+the compiled rule set from `.github/intent-policy.json` when present, otherwise the effective
+label-to-intent selectors derived from `.github/objective-mapping.json` — during validation
+or CI. Keys present in one source but not the other SHOULD surface as a sync warning or
+compliance failure so attribution and authorization stay aligned.
+
 ## Product boundary
 
 The system can establish:
diff --git a/specs/intent-attribution-compliance/README.md b/specs/intent-attribution-compliance/README.md
new file mode 100644
index 00000000000..1861b9462a1
--- /dev/null
+++ b/specs/intent-attribution-compliance/README.md
@@ -0,0 +1,26 @@
+# Intent Attribution Compliance Fixtures
+
+This directory defines the minimum compliance scenarios for the
+[Intent Attribution & Agent Governance Specification](../intent-attribution-agent-governance.md).
+
+## Required scenarios
+
+| Scenario | Expected result | Spec references |
+|---|---|---|
+| Explicit intent wins over linked issues | Attribution uses explicit metadata as the sole source | §RFC 2119 Norms → Attribution-Resolution Order |
+| Ambiguous root issue set | Attribution is `status: "ambiguous"` with `source: "closing_issue"` | §RFC 2119 Norms → Ambiguous-Root Handling |
+| Unlinked pull request fails closed | Governance resolves to the safest policy (`propose_only`, no writes, approval required) | §RFC 2119 Norms → Fail-Closed Behavior |
+
+## Fixture guidance
+
+Each future fixture in this directory SHOULD record:
+
+- the input artifact shape (explicit metadata, linked issues, labels)
+- the expected attribution source and status
+- the expected compiled execution policy when attribution is missing or ambiguous
+
+The minimum fixture set for conformance claims is:
+
+1. `explicit-intent-wins`
+2. `ambiguous-root-closing-issues`
+3. `unlinked-pr-fail-closed`
diff --git a/specs/otel-observability-spec.md b/specs/otel-observability-spec.md
index 990caf4fc2f..bbe4e2f38ab 100644
--- a/specs/otel-observability-spec.md
+++ b/specs/otel-observability-spec.md
@@ -863,6 +863,17 @@ The following test IDs are stubs for Level 1 (Stable Configuration and Export) c
 | **T-OT-006** | Local mirrors | The runtime helper writes each exported span as a raw OTLP/HTTP JSON line with a `resourceSpans` key to `/tmp/gh-aw/otel.jsonl`; the file format MUST NOT be an envelope-only summary. |
 | **T-OT-007** | Compiler config | `observability.otlp.headers` entries are emitted as `OTEL_EXPORTER_OTLP_HEADERS` in `key=value,key=value` format and are masked in diagnostics, job summaries, and artifacts. |
 
+### 17.1.2 Concrete Span-Attribute Contract Cases
+
+At minimum, the automated compliance suite SHOULD include the following
+attribute-level checks mapped directly to Section 10 requirements:
+
+| Test ID | Attribute requirement | Expected value | Verification method |
+|---------|------------------------|----------------|---------------------|
+| **T-OT-008** | §10.2 `gh-aw.job.name` on built-in job spans | The setup span includes `gh-aw.job.name` equal to the GitHub Actions job name | Decode `/tmp/gh-aw/otel.jsonl` or captured OTLP payloads and assert the setup span attribute value exactly matches the workflow job name |
+| **T-OT-009** | §10.3 `gen_ai.system` on built-in agent spans | A known engine emits a non-empty normalized provider/system value | Run a workflow with a built-in agent span, decode exported spans, and assert `gen_ai.system` is present whenever the engine mapping is known |
+| **T-OT-010** | §13.3 `gh-aw.outcome.type` on outcome-evaluation spans | Outcome-evaluation spans emit the safe-output type being evaluated | Execute the outcome collector, inspect the `gh-aw.outcome.evaluate` span, and assert `gh-aw.outcome.type` matches the evaluated manifest item type |
+
 
 ### 17.2 Optional Extension Tests
 
@@ -885,6 +896,10 @@ The following implementation areas are authoritative for version 0.4.0 compatibi
 
 Any change that alters a listed compatibility surface MUST update the corresponding tests in the same change.
 
+Outcome-evaluation span contracts SHOULD stay aligned with the governance and attribution
+schema defined in `specs/intent-attribution-agent-governance.md`, especially when intent
+context is added to outcome spans or links.
+
 ---
 
 ## 18. References
diff --git a/specs/replace-label-compliance/README.md b/specs/replace-label-compliance/README.md
index 232bc956ed3..c0fae81f5b9 100644
--- a/specs/replace-label-compliance/README.md
+++ b/specs/replace-label-compliance/README.md
@@ -12,6 +12,7 @@ requirements defined in §4 of the specification.
 | Filename | Scenario | Spec Coverage |
 |---|---|---|
 | `rl-001-glob-semantics.yaml` | Glob pattern matching for `allowed-add`, `allowed-remove`, and `blocked` follows gobwas/glob semantics | RL-001, T-RL-020–T-RL-023 |
+| `rl-002-allowlist-enforcement.yaml` | Non-empty allowlists enforce matches while empty allowlists permit any non-blocked label | RL-002, T-RL-021b, T-RL-022b, T-RL-025 |
 | `rl-003-blocklist-ordering.yaml` | Blocklist evaluation occurs before allowlist evaluation (security boundary) | RL-003, T-RL-023–T-RL-024 |
 
 ## Fixture Schema
@@ -57,6 +58,9 @@ The following test IDs defined in the replace-label specification map to these f
 |---------|---------|-------------|
 | T-RL-020 | `rl-001-glob-semantics.yaml` | Star glob matches label name substring |
 | T-RL-021 | `rl-001-glob-semantics.yaml` | Exact pattern matches only exact name |
+| T-RL-021b | `rl-002-allowlist-enforcement.yaml` | Non-empty allowed-add accepts a matching label |
 | T-RL-022 | `rl-001-glob-semantics.yaml` | Character class pattern matches correctly |
+| T-RL-022b | `rl-002-allowlist-enforcement.yaml` | Non-empty allowed-add rejects a non-matching label |
 | T-RL-023 | `rl-001-glob-semantics.yaml`, `rl-003-blocklist-ordering.yaml` | Glob pattern rejects non-matching label; blocked label rejected even when allowed |
 | T-RL-024 | `rl-003-blocklist-ordering.yaml` | Blocked label rejected even with wildcard allowed-add |
+| T-RL-025 | `rl-002-allowlist-enforcement.yaml` | Empty allowed-remove permits any non-blocked label |
diff --git a/specs/replace-label-compliance/rl-002-allowlist-enforcement.yaml b/specs/replace-label-compliance/rl-002-allowlist-enforcement.yaml
new file mode 100644
index 00000000000..722d4ca2782
--- /dev/null
+++ b/specs/replace-label-compliance/rl-002-allowlist-enforcement.yaml
@@ -0,0 +1,81 @@
+# RL-002 Allowlist Enforcement -- Compliance Fixture
+# Tests: T-RL-021, T-RL-022, T-RL-025
+# Spec: §4.1 Allowlist Enforcement (RL-002)
+
+fixture_id: "rl-002-allowlist-enforcement"
+description: >
+  When an allowlist is non-empty, a label name MUST match at least one configured
+  pattern before the operation is allowed. Empty or absent allowlists permit any
+  label name for that direction (RL-002).
+
+spec_refs:
+  - "RL-002 — Non-empty allowlists MUST enforce at least one matching pattern"
+  - "RL-002 — Empty allowlists permit any label name"
+  - "§4.1 — allowed-add and allowed-remove allowlist evaluation"
+  - "T-RL-021 — label_to_add matching allowed-add is accepted"
+  - "T-RL-022 — label_to_add not matching allowed-add is rejected"
+  - "T-RL-025 — label_to_remove is accepted when allowed-remove is empty"
+
+scenarios:
+  - scenario_id: "rl-002-allowed-add-match"
+    description: "A non-empty allowed-add list permits an exact label match."
+    input:
+      safe_output_config:
+        allowed-add:
+          - "done"
+        allowed-remove: []
+        blocked: []
+      message:
+        label_to_add: "done"
+        label_to_remove: ""
+    expected:
+      decision: allow
+      error_code: null
+      reason: ""
+
+  - scenario_id: "rl-002-allowed-add-miss"
+    description: "A non-empty allowed-add list rejects a label that matches no pattern."
+    input:
+      safe_output_config:
+        allowed-add:
+          - "done"
+        allowed-remove: []
+        blocked: []
+      message:
+        label_to_add: "wontfix"
+        label_to_remove: ""
+    expected:
+      decision: deny
+      error_code: -32002
+      reason: "label not in allowed-add list"
+
+  - scenario_id: "rl-002-empty-allowed-remove-permits"
+    description: "An empty allowed-remove list permits removing any non-blocked label."
+    input:
+      safe_output_config:
+        allowed-add: []
+        allowed-remove: []
+        blocked: []
+      message:
+        label_to_add: ""
+        label_to_remove: "state-in-progress"
+    expected:
+      decision: allow
+      error_code: null
+      reason: ""
+
+  - scenario_id: "rl-002-allowed-remove-miss"
+    description: "A non-empty allowed-remove list rejects removal when no pattern matches."
+    input:
+      safe_output_config:
+        allowed-add: []
+        allowed-remove:
+          - "state-*"
+        blocked: []
+      message:
+        label_to_add: ""
+        label_to_remove: "bug"
+    expected:
+      decision: deny
+      error_code: -32002
+      reason: "label not in allowed-remove list"
diff --git a/specs/replace-label-spec.md b/specs/replace-label-spec.md
index c67c57ff097..8d12683d714 100644
--- a/specs/replace-label-spec.md
+++ b/specs/replace-label-spec.md
@@ -534,6 +534,10 @@ The test suite for `replace-label` spans two layers:
 
 #### 9.2.3 Label Validation Tests
 
+The normative compliance fixtures for the allowlist and blocklist edge cases in
+this subsection live in `specs/replace-label-compliance/rl-002-allowlist-enforcement.yaml`
+and `specs/replace-label-compliance/rl-003-blocklist-ordering.yaml`.
+
 - **T-RL-020**: Verify that `label_to_add` is accepted when `allowed-add` is empty.
 - **T-RL-021**: Verify that `label_to_add` matching a pattern in `allowed-add` is accepted.
 - **T-RL-022**: Verify that `label_to_add` not matching any pattern in a non-empty `allowed-add` is rejected.
diff --git a/specs/safe-output-outcome-evaluation.md b/specs/safe-output-outcome-evaluation.md
index 17a825bacad..73351a2748e 100644
--- a/specs/safe-output-outcome-evaluation.md
+++ b/specs/safe-output-outcome-evaluation.md
@@ -111,7 +111,7 @@ Rows marked `evalGenericSticky` fallback are generic existence checks, not type-
 | `link_sub_issue` | `evalGenericSticky` fallback | sub-issue link target exists |
 | `hide_comment` | `evalHideComment` | none yet |
 | `assign_milestone` | `evalAssignMilestone` | milestone still set |
-| `replace_label` | `evalGenericSticky` fallback | label target still exists |
+| `replace_label` | `evalReplaceLabel` | label replacement retained |
 | `update_project` | `evalGenericSticky` fallback | object still exists |
 | `update_release` | `evalGenericSticky` fallback | object still exists |
 | `noop` | explicit skip | skipped |
@@ -122,12 +122,12 @@ Rows marked `evalGenericSticky` fallback are generic existence checks, not type-
 | `create_pull_request` | implemented | `pkg/workflow/safe_outputs_config.go`, `pkg/workflow/compiler_safe_outputs.go`, `pkg/cli/outcome_eval.go` (`evalCreatePullRequest`) | `actions/setup/js/safe_outputs_handlers.cjs`, `actions/setup/js/evaluate_outcomes.cjs` (PR-specific path) |
 | `create_issue` | implemented | `pkg/workflow/safe_outputs_config.go`, `pkg/workflow/compiler_safe_outputs.go`, `pkg/cli/outcome_eval.go` (`evalCreateIssue`) | `actions/setup/js/safe_outputs_handlers.cjs`, `actions/setup/js/evaluate_outcomes.cjs` (issue-specific path) |
 | `add_comment` | implemented | `pkg/workflow/safe_outputs_dispatch.go`, `pkg/cli/outcome_eval.go` (`evalAddComment`) | `actions/setup/js/safe_outputs_handlers.cjs`, `actions/setup/js/evaluate_outcomes.cjs` (issue-comment URL path) |
-| `add_labels` | partial | `pkg/workflow/safe_outputs_allowed_labels_validation.go`, `pkg/cli/outcome_eval.go` (`evalAddLabels`) | `actions/setup/js/safe_outputs_handlers.cjs`, `actions/setup/js/evaluate_outcomes.cjs` (generic fallback) |
+| `add_labels` | partial | `pkg/workflow/safe_outputs_allowed_labels_validation.go`, `pkg/cli/outcome_eval.go` (`evalAddLabels`) | `actions/setup/js/safe_outputs_handlers.cjs`, `actions/setup/js/evaluate_outcomes.cjs` (`evaluateAddLabels`) |
 | `add_reviewer` | implemented | `pkg/workflow/add_reviewer.go`, `pkg/workflow/safe_outputs_config.go`, `pkg/cli/outcome_eval_review.go` | `actions/setup/js/add_reviewer.cjs`, `actions/setup/js/evaluate_outcomes.cjs` |
 | `update_issue` | implemented | `pkg/workflow/safe_outputs_config.go`, `pkg/cli/outcome_eval_update.go` | `actions/setup/js/update_issue.cjs`, `actions/setup/js/evaluate_outcomes.cjs` |
 | `update_pull_request` | implemented | `pkg/workflow/safe_outputs_config.go`, `pkg/cli/outcome_eval_update.go` | `actions/setup/js/update_pull_request.cjs`, `actions/setup/js/evaluate_outcomes.cjs` |
-| `close_issue` | partial | `pkg/workflow/safe_outputs_dispatch.go`, `pkg/cli/outcome_eval.go` (`evalCloseSticky`) | `actions/setup/js/close_issue.cjs`, `actions/setup/js/evaluate_outcomes.cjs` (generic fallback) |
-| `close_pull_request` | partial | `pkg/workflow/safe_outputs_dispatch.go`, `pkg/cli/outcome_eval.go` (`evalCloseSticky`) | `actions/setup/js/close_pull_request.cjs`, `actions/setup/js/evaluate_outcomes.cjs` (generic fallback) |
+| `close_issue` | partial | `pkg/workflow/safe_outputs_dispatch.go`, `pkg/cli/outcome_eval.go` (`evalCloseSticky`) | `actions/setup/js/close_issue.cjs`, `actions/setup/js/evaluate_outcomes.cjs` (`evaluateCloseIssue`) |
+| `close_pull_request` | partial | `pkg/workflow/safe_outputs_dispatch.go`, `pkg/cli/outcome_eval.go` (`evalCloseSticky`) | `actions/setup/js/close_pull_request.cjs`, `actions/setup/js/evaluate_outcomes.cjs` (`evaluateClosePullRequest`) |
 | `close_discussion` | partial | `pkg/workflow/safe_outputs_dispatch.go`, `pkg/cli/outcome_eval.go` (`evalCloseDiscussion`) | `actions/setup/js/close_discussion.cjs`, `actions/setup/js/evaluate_outcomes.cjs` (generic fallback) |
 | `create_discussion` | partial | `pkg/workflow/safe_outputs_dispatch.go`, `pkg/cli/outcome_eval.go` (`evalCreateDiscussion`) | `actions/setup/js/create_discussion.cjs`, `actions/setup/js/evaluate_outcomes.cjs` (generic fallback) |
 | `update_discussion` | partial | `pkg/workflow/safe_outputs_config.go`, `pkg/cli/outcome_eval_workflow.go` (`evalUpdateDiscussion`) | `actions/setup/js/update_discussion.cjs`, `actions/setup/js/evaluate_outcomes.cjs` (generic fallback) |
@@ -144,7 +144,7 @@ Rows marked `evalGenericSticky` fallback are generic existence checks, not type-
 | `link_sub_issue` | not-started | `pkg/workflow/link_sub_issue.go`, `pkg/cli/outcome_eval.go` (`evalGenericSticky` fallback) | `actions/setup/js/link_sub_issue.cjs`, `actions/setup/js/evaluate_outcomes.cjs` (generic fallback) |
 | `hide_comment` | partial | `pkg/workflow/hide_comment.go`, `pkg/cli/outcome_eval.go` (`evalHideComment`) | `actions/setup/js/hide_comment.cjs`, `actions/setup/js/evaluate_outcomes.cjs` (generic fallback) |
 | `assign_milestone` | partial | `pkg/workflow/assign_milestone.go`, `pkg/cli/outcome_eval.go` (`evalAssignMilestone`) | `actions/setup/js/assign_milestone.cjs`, `actions/setup/js/evaluate_outcomes.cjs` (generic fallback) |
-| `replace_label` | not-started | `pkg/workflow/replace_label.go`, `pkg/cli/outcome_eval.go` (`evalGenericSticky` fallback) | `actions/setup/js/replace_label.cjs`, `actions/setup/js/evaluate_outcomes.cjs` (generic fallback) |
+| `replace_label` | partial | `pkg/workflow/replace_label.go`, `pkg/cli/outcome_eval.go` (`evalReplaceLabel`) | `actions/setup/js/replace_label.cjs`, `actions/setup/js/evaluate_outcomes.cjs` (generic fallback) |
 | `update_project` | not-started | `pkg/workflow/update_project.go`, `pkg/cli/outcome_eval.go` (`evalGenericSticky` fallback) | `actions/setup/js/update_project.cjs`, `actions/setup/js/evaluate_outcomes.cjs` (generic fallback) |
 | `update_release` | not-started | `pkg/workflow/safe_outputs_config.go`, `pkg/cli/outcome_eval.go` (`evalGenericSticky` fallback) | `actions/setup/js/update_release.cjs`, `actions/setup/js/evaluate_outcomes.cjs` (generic fallback) |
 | `noop` | implemented | `pkg/cli/outcome_eval.go` (explicit skip in `EvaluateOutcomes`) | `actions/setup/js/evaluate_outcomes.cjs` (`NOOP_TYPES`) |