fix(workflow): forward-port safe-output reliability hardening from #1683

test/gh-aw-quality.test.ts

@@ -238,6 +238,12 @@ describe('gh-aw: safe-output configuration', () => {
     expect(safeOutputs['create-issue'], 'create-issue should exist').toBeDefined();
     expect(safeOutputs['add-comment'], 'add-comment should exist').toBeDefined();
   });
+
+  it('create-issue max is 75 (supports large plans, forward-port of #1683)', () => {
+    const ci = safeOutputs['create-issue'];
+    expect(ci, 'create-issue block must exist').toBeDefined();
+    expect(ci['max'], 'create-issue max must be 75 — do not reduce below this').toBe(75);
+  });
 });
 
 // ---------------------------------------------------------------------------
@@ -528,4 +534,63 @@ describe('gh-aw: prompt budget & planning import regression', () => {
       `Combined prompt is ${totalKB} KB — exceeds the gh-aw ${GH_AW_PROMPT_CEILING_KB} KB ceiling. Headroom: ${headroomKB} KB.`
     ).toBeLessThan(GH_AW_PROMPT_CEILING_BYTES);
   });
+
+  it('reports combined bytes and headroom', () => {
+    let totalBytes = Buffer.byteLength(squadContent, 'utf8');
+    for (const importPath of imports) {
+      const fullPath = join(WORKFLOWS_DIR, importPath);
+      if (existsSync(fullPath)) totalBytes += Buffer.byteLength(readFileSync(fullPath, 'utf8'), 'utf8');
+    }
+    const headroomBytes = GH_AW_PROMPT_CEILING_BYTES - totalBytes;
+    // Informational — log bytes/headroom; fail only if headroom < 5 KB (regression guard)
+    expect(
+      headroomBytes,
+      `Headroom too low: ${(headroomBytes / 1024).toFixed(1)} KB remaining of ${GH_AW_PROMPT_CEILING_KB} KB ceiling`
+    ).toBeGreaterThan(5 * 1024);
+  });
+});
+
+// ---------------------------------------------------------------------------
+// Test: Plan Activate hardening behaviors (forward-port #1683)
+// ---------------------------------------------------------------------------
+
+describe('gh-aw: Plan Activate hardening behaviors', () => {
+  const content = readFileSync(SQUAD_WORKFLOW, 'utf8');
+
+  it('includes output budget awareness guidance', () => {
+    expect(content).toContain('Output Budget Awareness');
+    expect(content).toMatch(/total.*>\s*50.*phased activation|phased.*activation.*>\s*50/i);
+  });
+
+  it('includes label pre-flight step before issue creation', () => {
+    expect(content).toContain('Label Pre-flight');
+    expect(content).toMatch(/squad.*label.*exist|label.*squad.*exist/i);
+  });
+
+  it('label pre-flight does not claim impossible label creation', () => {
+    // Workflow has issues: read and no create-label safe-output; must not claim it can create labels
+    const preflight = content.match(/##### Label Pre-flight\n([\s\S]*?)(?=\n#####|\n####)/)?.[1] ?? '';
+    expect(preflight).not.toMatch(/create them|safe-output permissions handle the write|no additional token scope/i);
+  });
+
+  it('label pre-flight reports missing labels as prerequisite gap', () => {
+    expect(content).toMatch(/prerequisite gap|prerequisite/i);
+    expect(content).toMatch(/issues: write/i);
+    expect(content).toMatch(/create-label.*safe-output|safe-output.*create-label/i);
+  });
+
+  it('label pre-flight continues activation and omits unavailable labels with report', () => {
+    expect(content).toMatch(/unavailable labels are omitted and reported|omitted and reported/i);
+  });
+
+  it('includes transient failure handling with single retry', () => {
+    expect(content).toContain('Transient Failure Handling');
+    expect(content).toMatch(/5xx.*retry|retry.*5xx/i);
+  });
+
+  it('includes graceful sub-issue fallback for 404/422', () => {
+    expect(content).toContain('Sub-issue Fallback');
+    expect(content).toMatch(/404.*422|422.*404/);
+    expect(content).toMatch(/degrade gracefully|graceful/i);
+  });
 });

workflows/squad.md

@@ -50,7 +50,7 @@ safe-outputs:
     expires: 14d
   create-issue:
     labels: [squad]
-    max: 50
+    max: 75
   add-comment:
     max: 20
 ---
@@ -711,6 +711,22 @@ After phase acceptance, check if ready for automatic activation:
 
 After EVERY `create-issue` call: verify returned issue number, stop on failure, NEVER predict issue numbers.
 
+##### Output Budget Awareness
+
+Count expected issues before starting. If total > 50: recommend phased activation (`/squad plan activate phase {N}`) and proceed with the current phase only. If total > 30: use compact issue bodies (scope + acceptance criteria only; omit elaboration).
+
+##### Label Pre-flight
+
+Before the first `create-issue`, verify labels `squad` and any `squad:{agent}` exist. If missing, record them in the activation summary as a prerequisite gap (label creation requires `issues: write` + `create-label` safe-output — not configured in this workflow). Continue activation — `create-issue` will apply any existing labels normally; unavailable labels are omitted and reported, not silently applied.
+
+##### Transient Failure Handling
+
+On `5xx` response from `create-issue`: wait briefly and retry once. On second failure or `4xx`: record the issue title as skipped in the activation summary, continue with remaining issues. Never abort the full run for a single transient failure.
+
+##### Sub-issue Fallback
+
+When setting a `parent` sub-issue relationship returns `404` or `422` (feature disabled or repo plan): degrade gracefully — record the intended parent as a body reference (`Parent: #{issue_number}`), then continue. Never fail activation over sub-issue API unavailability.
+
 ##### Step 2: Create Issues — Full Hierarchy
 
 Root → Epics → Tasks. Phase-specific: filter to matching phase heading.