@boyingliu01/xp-gate 0.35.3 → 0.35.5

package/lib/mattpocock-license-verdict.json

@@ -2,6 +2,6 @@
   "verdict": "PASS",
   "license_type": "MIT",
   "sha256": "0e7ac423bf2c6e223b7c5b156f8cf72da49d748e56a1641402c31f22ad07dbb5",
-  "checked_at": "2026-10-30T16:54:54.104Z",
+  "checked_at": "2026-10-30T18:00:25.743Z",
   "source": "network"
 }

package/lib/retro.js

@@ -14,6 +14,7 @@
 const fs = require('fs');
 const path = require('path');
 const { execSync } = require('child_process');
+const { computeFlakyStats, parseQuarantine, findPolicyViolations } = require('./flaky');
 
 /**
  * Upper bound for a believable gate runtime. A gate that reports more than two
@@ -276,6 +277,31 @@ function buildReport(opts) {
   }
   L.push('');
 
+  // Flaky-test governance (#603)
+  L.push('## Test Stability (#603)');
+  const flaky = computeFlakyStats(gateEntries);
+  if (!flaky.measurable) {
+    // Say "not measured" rather than "0%". A journal with no re-runs cannot
+    // distinguish a stable suite from an unmeasured one, and printing 0% would
+    // claim the former.
+    L.push(`⚠️  NOT MEASURABLE — ${flaky.reason}`);
+  } else {
+    L.push(`- First-pass rate: ${flaky.firstPassRate}% (${flaky.firstPasses}/${flaky.firstAttempts} first attempts)`);
+    L.push(`- Flaky recoveries: ${flaky.flakyRecoveries} of ${flaky.totalRuns} run(s) passed only on re-run (${flaky.flakyRate}%)`);
+    // The pass rate above counts re-run successes, so state the gap explicitly.
+    L.push(`- Note: the gate pass rate above counts re-run successes; first-pass rate is the honest number.`);
+  }
+  const quarantineEntries = parseQuarantine(path.join(process.cwd(), '.flaky-quarantine.yaml'));
+  if (quarantineEntries.length > 0) {
+    const violations = findPolicyViolations(quarantineEntries, now);
+    L.push(`- Quarantined tests: ${quarantineEntries.length} (${violations.length} policy violation(s))`);
+    for (const v of violations) {
+      // An expired entry is a permanent deletion wearing a quarantine badge.
+      L.push(`  - ${v.rule}: ${v.entry.test} (owner ${v.entry.owner || 'unassigned'}, expiry ${v.entry.expiry || 'none'})`);
+    }
+  }
+  L.push('');
+
   // Rework Rate Trend (#369)
   L.push('## Rework Rate Trend (#369)');
   const withRate = sprints.filter(s => s.metrics && s.metrics.rework_rate != null);

package/mock-policy/AGENTS.md

@@ -1,9 +1,9 @@
 # SRC/MOCK-POLICY KNOWLEDGE BASE
 
 **Generated:** 2026-10-30
-**Commit:** 9ebd1048
+**Commit:** 4776dc7f
 **Branch:** main
-**Version:** 0.35.3.0
+**Version:** 0.35.5.0
 
 ## OVERVIEW
 Mock layering policy enforcement — Gate M3 of pre-push hook. Ensures integration tests use real implementations for internal dependencies, mock external dependencies, and annotate pending mocks with removal plans. Combines project scope scanning, mock decision engine, and per-file validation into a single pipeline.

package/mutation/AGENTS.md

@@ -1,9 +1,9 @@
 # SRC/MUTATION KNOWLEDGE BASE
 
 **Generated:** 2026-10-30
-**Commit:** 9ebd1048
+**Commit:** 4776dc7f
 **Branch:** main
-**Version:** 0.35.3.0
+**Version:** 0.35.5.0
 
 ## OVERVIEW
 **Gate M** (incremental mutation testing) + **Gate M2** helpers (test-layer detection used by `src/mock-policy/`). Pre-push quality gate. TypeScript-only; uses Stryker.

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@boyingliu01/xp-gate",
-  "version": "0.35.3",
+  "version": "0.35.5",
   "description": "AI-driven development workflow: 6 quality gates + Delphi review + Sprint Flow",
   "bin": {
     "xp-gate": "bin/xp-gate.js"

package/plugins/claude-code/.claude-plugin/plugin.json

@@ -1,6 +1,6 @@
 {
   "name": "xp-gate",
-  "version": "0.35.3",
+  "version": "0.35.5",
   "displayName": "XP-Gate",
   "description": "Extreme Programming quality gates + AI workflow skills for Claude Code. Includes 12 pre-commit gates (Gate 0-11), 8 pre-push gates, Sprint Flow (6 phases), and Delphi multi-expert review (>=90% consensus).",
   "author": {

package/plugins/claude-code/skills/delphi-review/AGENTS.md

@@ -1,9 +1,9 @@
 # SKILLS/DELPHI-REVIEW KNOWLEDGE BASE
 
 **Generated:** 2026-10-30
-**Commit:** 9ebd1048
+**Commit:** 4776dc7f
 **Branch:** main
-**Version:** 0.35.3.0
+**Version:** 0.35.5.0
 
 ## OVERVIEW
 Delphi Consensus Review — multi-round anonymous expert review (≥90% threshold, exactly 3 experts with distinct executable model IDs). Supports design, requirements, and code-walkthrough modes.

package/plugins/claude-code/skills/sprint-flow/AGENTS.md

@@ -1,9 +1,9 @@
 # SKILLS/SPRINT-FLOW KNOWLEDGE BASE
 
 **Generated:** 2026-10-30
-**Commit:** 9ebd1048
+**Commit:** 4776dc7f
 **Branch:** main
-**Version:** 0.35.3.0
+**Version:** 0.35.5.0
 
 ## OVERVIEW
 **6-phase** development pipeline (v2.0 compact redesign, Issue #290): PREP → DESIGN → BUILD → VERIFY → SHIP → CLOSE. Phase 3/6 BUILD default build mode is **ralph-loop** (REQ-level iteration, 40-67% token savings vs parallel). HARD-GATE between DESIGN (2/6) and BUILD (3/6): design must pass Delphi review (≥90% consensus) before any coding.

package/plugins/claude-code/skills/test-specification-alignment/AGENTS.md

@@ -1,9 +1,9 @@
 # SKILLS/TEST-SPECIFICATION-ALIGNMENT KNOWLEDGE BASE
 
 **Generated:** 2026-10-30
-**Commit:** 9ebd1048
+**Commit:** 4776dc7f
 **Branch:** main
-**Version:** 0.35.3.0
+**Version:** 0.35.5.0
 
 ## OVERVIEW
 Test-Specification Alignment Engine — two-stage validation ensuring tests accurately reflect requirements and design specs.

package/plugins/dsh/lib/delphi-run.d.ts

@@ -19,6 +19,5 @@ export interface ReviewOutcome {
  * plus whatever the caller says was addressed, then re-votes.
  */
 export declare function runReview(opts: DelphiOptions, subject: string, addressed?: string, signal?: AbortSignal): Promise<ReviewOutcome>;
-/** Human-readable transcript for the tool result. */
 export declare function formatOutcome(outcome: ReviewOutcome): string;
 //# sourceMappingURL=delphi-run.d.ts.map

package/plugins/dsh/lib/delphi-run.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"delphi-run.d.ts","sourceRoot":"","sources":["../src/delphi-run.ts"],"names":[],"mappings":"AAAA;;;;;;;;GAQG;AAKH,OAAO,KAAK,EAAE,aAAa,EAAc,WAAW,EAAE,MAAM,aAAa,CAAA;AAoBzE,MAAM,WAAW,aAAa;IAC5B,QAAQ,CAAC,MAAM,EAAE,SAAS,WAAW,EAAE,CAAA;IACvC,QAAQ,CAAC,SAAS,EAAE,OAAO,CAAA;IAC3B,QAAQ,CAAC,UAAU,EAAE,MAAM,CAAA;IAC3B,QAAQ,CAAC,WAAW,EAAE,MAAM,CAAA;CAC7B;AA0CD;;;GAGG;AACH,wBAAsB,SAAS,CAC7B,IAAI,EAAE,aAAa,EACnB,OAAO,EAAE,MAAM,EACf,SAAS,SAAK,EACd,MAAM,CAAC,EAAE,WAAW,GACnB,OAAO,CAAC,aAAa,CAAC,CA0CxB;AAED,qDAAqD;AACrD,wBAAgB,aAAa,CAAC,OAAO,EAAE,aAAa,GAAG,MAAM,CAiB5D"}
+{"version":3,"file":"delphi-run.d.ts","sourceRoot":"","sources":["../src/delphi-run.ts"],"names":[],"mappings":"AAAA;;;;;;;;GAQG;AAKH,OAAO,KAAK,EAAE,aAAa,EAA6B,WAAW,EAAE,MAAM,aAAa,CAAA;AAoBxF,MAAM,WAAW,aAAa;IAC5B,QAAQ,CAAC,MAAM,EAAE,SAAS,WAAW,EAAE,CAAA;IACvC,QAAQ,CAAC,SAAS,EAAE,OAAO,CAAA;IAC3B,QAAQ,CAAC,UAAU,EAAE,MAAM,CAAA;IAC3B,QAAQ,CAAC,WAAW,EAAE,MAAM,CAAA;CAC7B;AA0CD;;;GAGG;AACH,wBAAsB,SAAS,CAC7B,IAAI,EAAE,aAAa,EACnB,OAAO,EAAE,MAAM,EACf,SAAS,SAAK,EACd,MAAM,CAAC,EAAE,WAAW,GACnB,OAAO,CAAC,aAAa,CAAC,CA0CxB;AAoDD,wBAAgB,aAAa,CAAC,OAAO,EAAE,aAAa,GAAG,MAAM,CAe5D"}

package/plugins/dsh/lib/delphi-run.js

@@ -94,15 +94,65 @@ export async function runReview(opts, subject, addressed = "", signal) {
     };
 }
 /** Human-readable transcript for the tool result. */
+/**
+ * Reduce a seat's raw report to the part worth reading (#670).
+ *
+ * The seat was asked for `### Findings` / `### VERDICT` / `### CONFIDENCE`; the
+ * verdict and confidence are already on the tally line above, so repeating them
+ * here only makes the block longer. What the caller cannot get anywhere else is
+ * the Findings table, which is why the report is trimmed down to it.
+ */
+function findingsExcerpt(report) {
+    const text = report.trim();
+    if (text === "")
+        return "";
+    // Split on the section headings the prompt asks for; no regex escaping needed,
+    // which keeps this readable and avoids a silently-mangled character class.
+    const lines = text.split(/\r?\n/);
+    const start = lines.findIndex((line) => /^#{1,3}\s*Findings\s*$/i.test(line));
+    if (start === -1) {
+        // No Findings heading: keep everything before VERDICT rather than returning
+        // nothing on an unexpected format.
+        const stop = lines.findIndex((line) => /^#{1,3}\s*VERDICT/i.test(line));
+        return (stop === -1 ? lines : lines.slice(0, stop)).join("\n").trim();
+    }
+    let end = lines.length;
+    for (let i = start + 1; i < lines.length; i += 1) {
+        if (/^#{1,3}\s*[A-Z]/i.test(lines[i])) {
+            end = i;
+            break;
+        }
+    }
+    return lines.slice(start + 1, end).join("\n").trim();
+}
+/**
+ * One seat's tally line, plus its findings when this is the final round (#670).
+ *
+ * Only the final round expands: printing every round's full report would bury the
+ * conclusion under five historical copies, and what the caller needs is "what is
+ * still wrong", not a transcript.
+ */
+function verdictLines(v, expand) {
+    const model = v.resolvedModel ? `${v.requestedModel} -> ${v.resolvedModel}` : v.requestedModel;
+    const note = v.error ? ` (${v.error})` : "";
+    const lines = [`- [${v.role}] ${v.id} (${model}): ${v.verdict} confidence=${v.confidence}${note}`];
+    if (!expand)
+        return lines;
+    const excerpt = findingsExcerpt(v.report);
+    if (excerpt === "")
+        return lines;
+    for (const line of excerpt.split(/\r?\n/))
+        lines.push(`    ${line}`);
+    return lines;
+}
 export function formatOutcome(outcome) {
     const lines = [];
+    const lastRound = outcome.rounds.length;
     for (const r of outcome.rounds) {
         lines.push(`## Round ${r.round} — approvals ${r.approvals}/${r.validCount} valid`);
-        for (const v of r.verdicts) {
-            const model = v.resolvedModel ? `${v.requestedModel} -> ${v.resolvedModel}` : v.requestedModel;
-            const note = v.error ? ` (${v.error})` : "";
-            lines.push(`- [${v.role}] ${v.id} (${model}): ${v.verdict} confidence=${v.confidence}${note}`);
-        }
+        const expand = r.round === lastRound;
+        for (const v of r.verdicts)
+            lines.push(...verdictLines(v, expand));
         lines.push("");
     }
     lines.push(outcome.consensus

package/plugins/dsh/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@boyingliu01/dsh-plugin-xp-gate",
-  "version": "0.35.3",
+  "version": "0.35.5",
   "description": "XP-Gate deterministic quality gates + AI workflow skills as a DeepSeek Harness (DSH) native plugin: gate-check / gate-principles / gate-arch tools + 12 SKILL.md skills",
   "type": "module",
   "main": "./lib/index.js",

package/plugins/dsh/skills/delphi-review/AGENTS.md

@@ -1,9 +1,9 @@
 # SKILLS/DELPHI-REVIEW KNOWLEDGE BASE
 
 **Generated:** 2026-10-30
-**Commit:** 9ebd1048
+**Commit:** 4776dc7f
 **Branch:** main
-**Version:** 0.35.3.0
+**Version:** 0.35.5.0
 
 ## OVERVIEW
 Delphi Consensus Review — multi-round anonymous expert review (≥90% threshold, exactly 3 experts with distinct executable model IDs). Supports design, requirements, and code-walkthrough modes.

package/plugins/dsh/skills/sprint-flow/AGENTS.md

@@ -1,9 +1,9 @@
 # SKILLS/SPRINT-FLOW KNOWLEDGE BASE
 
 **Generated:** 2026-10-30
-**Commit:** 9ebd1048
+**Commit:** 4776dc7f
 **Branch:** main
-**Version:** 0.35.3.0
+**Version:** 0.35.5.0
 
 ## OVERVIEW
 **6-phase** development pipeline (v2.0 compact redesign, Issue #290): PREP → DESIGN → BUILD → VERIFY → SHIP → CLOSE. Phase 3/6 BUILD default build mode is **ralph-loop** (REQ-level iteration, 40-67% token savings vs parallel). HARD-GATE between DESIGN (2/6) and BUILD (3/6): design must pass Delphi review (≥90% consensus) before any coding.

package/plugins/dsh/skills/test-specification-alignment/AGENTS.md

@@ -1,9 +1,9 @@
 # SKILLS/TEST-SPECIFICATION-ALIGNMENT KNOWLEDGE BASE
 
 **Generated:** 2026-10-30
-**Commit:** 9ebd1048
+**Commit:** 4776dc7f
 **Branch:** main
-**Version:** 0.35.3.0
+**Version:** 0.35.5.0
 
 ## OVERVIEW
 Test-Specification Alignment Engine — two-stage validation ensuring tests accurately reflect requirements and design specs.

package/plugins/dsh/src/delphi-findings.test.ts

@@ -0,0 +1,109 @@
+import { describe, it, expect } from "vitest"
+import { formatOutcome } from "./delphi-run.js"
+import type { ReviewOutcome } from "./delphi-run.js"
+
+/**
+ * @test REQ-670 delphi 报告必须回传专家意见正文
+ * @intent `formatOutcome` 只回传 verdict 与 confidence，把专家写的
+ *         `### Findings` 表格整块丢掉。调用方拿到的是一串 REQUEST_CHANGES
+ *         加上 0 条具体问题——评审消耗了真实模型调用，产出的却是不可行动的信息。
+ *
+ *         这个 bug 有个特殊之处：**数据一直都在**。`ExpertVerdict.report` 完整
+ *         保留了专家原文，只是呈现层从来没读过它。所以这不是"信息不足"，是
+ *         "攒了信息然后扔掉"——比前者更难发现，因为所有中间层看起来都是对的。
+ *
+ *         修复必须小心一个副作用：报告是多轮匿名评审，专家在后续轮次会看到
+ *         聚合票型。若把每轮全文都打印出来，输出会迅速膨胀到不可读（3 席 × 5 轮
+ *         × 300 字）。所以正文只在**最后一轮**展开，前面各轮仍只给票型——读者
+ *         需要的是"最终为什么不通过"，不是五份历史副本。
+ * @covers AC-670-66, AC-670-67
+ */
+
+const FINDINGS_REPORT =
+  "### Findings\n| ID | Severity | Description | Location | Fix |\n|---|---|---|---|---|\n" +
+  "| F1 | critical | exit code discarded | run.ts:42 | propagate it |\n" +
+  "\n### VERDICT\nREQUEST_CHANGES\n### CONFIDENCE\n9"
+
+const APPROVED_REPORT = "### Findings\nNONE\n\n### VERDICT\nAPPROVED\n### CONFIDENCE\n8"
+
+function verdict(id: string, role: string, report: string, verdictValue: string, confidence: number) {
+  return {
+    id,
+    role,
+    requestedModel: `m-${id}`,
+    resolvedModel: `res-${id}`,
+    verdict: verdictValue as "APPROVED" | "REQUEST_CHANGES",
+    confidence,
+    counted: true,
+    report,
+  }
+}
+
+function outcomeOf(rounds: Array<{ round: number; verdicts: ReturnType<typeof verdict>[] }>): ReviewOutcome {
+  const mapped = rounds.map((r) => ({
+    round: r.round,
+    approvals: r.verdicts.filter((v) => v.verdict === "APPROVED").length,
+    validCount: r.verdicts.length,
+    verdicts: r.verdicts,
+  }))
+  return {
+    rounds: mapped,
+    consensus: false,
+    finalRatio: 0,
+    totalRounds: mapped.length,
+  }
+}
+
+describe("formatOutcome 回传专家意见 (#670)", () => {
+  it("AC-670-66: 最终轮的专家意见正文必须出现在输出里", () => {
+    const out = outcomeOf([
+      {
+        round: 1,
+        verdicts: [
+          verdict("A", "architecture", FINDINGS_REPORT, "REQUEST_CHANGES", 9),
+          verdict("B", "technical", APPROVED_REPORT, "APPROVED", 8),
+        ],
+      },
+    ])
+    const text = formatOutcome(out)
+    // The whole point: a reviewer must be able to read what was wrong.
+    expect(text).toContain("exit code discarded")
+    expect(text).toContain("propagate it")
+    // The verdict/confidence are already on the tally line; the excerpt must not
+    // repeat them, or every seat's block doubles in size.
+    expect(text).not.toContain("### VERDICT")
+  })
+
+  it("AC-670-67: 中间轮次只给票型，不重复整段正文（否则输出不可读）", () => {
+    const out = outcomeOf([
+      {
+        round: 1,
+        verdicts: [verdict("A", "architecture", FINDINGS_REPORT, "REQUEST_CHANGES", 9)],
+      },
+      {
+        round: 2,
+        verdicts: [verdict("A", "architecture", APPROVED_REPORT, "APPROVED", 8)],
+      },
+    ])
+    const text = formatOutcome(out)
+    // The final round approved with NONE, so the earlier rejection's findings are
+    // historical and must not be expanded -- otherwise the reader sees a problem
+    // that was already resolved as if it were still open.
+    const occurrences = text.split("exit code discarded").length - 1
+    expect(occurrences).toBe(0)
+    // But the earlier round's tally line is still there: history is not erased.
+    expect(text).toContain("## Round 1")
+    expect(text).toContain("REQUEST_CHANGES")
+  })
+
+  it("AC-670-66: 报告无 Findings 标题时仍保留正文，而不是整块丢弃", () => {
+    // The original bug's shape was "unrecognized format -> print nothing". A
+    // fallback that also prints nothing would reproduce it under a new input.
+    const freeform = "The suffix resets $? so every gate reads as passing.\n\n### VERDICT\nREQUEST_CHANGES\n### CONFIDENCE\n9"
+    const out = outcomeOf([
+      { round: 1, verdicts: [verdict("A", "architecture", freeform, "REQUEST_CHANGES", 9)] },
+    ])
+    const text = formatOutcome(out)
+    expect(text).toContain("resets $?")
+  })
+})

package/plugins/dsh/src/delphi-run.ts

@@ -11,7 +11,7 @@
 import { TOOL_ABORTED } from "@deepseek-ai/dsh-tools"
 import { HarnessError } from "@deepseek-ai/dsh-llm"
 
-import type { DelphiOptions, ExpertSeat, RoundResult } from "./delphi.js"
+import type { DelphiOptions, ExpertSeat, ExpertVerdict, RoundResult } from "./delphi.js"
 import { runRound } from "./delphi.js"
 
 /**
@@ -132,15 +132,62 @@ export async function runReview(
 }
 
 /** Human-readable transcript for the tool result. */
+/**
+ * Reduce a seat's raw report to the part worth reading (#670).
+ *
+ * The seat was asked for `### Findings` / `### VERDICT` / `### CONFIDENCE`; the
+ * verdict and confidence are already on the tally line above, so repeating them
+ * here only makes the block longer. What the caller cannot get anywhere else is
+ * the Findings table, which is why the report is trimmed down to it.
+ */
+function findingsExcerpt(report: string): string {
+  const text = report.trim()
+  if (text === "") return ""
+  // Split on the section headings the prompt asks for; no regex escaping needed,
+  // which keeps this readable and avoids a silently-mangled character class.
+  const lines = text.split(/\r?\n/)
+  const start = lines.findIndex((line) => /^#{1,3}\s*Findings\s*$/i.test(line))
+  if (start === -1) {
+    // No Findings heading: keep everything before VERDICT rather than returning
+    // nothing on an unexpected format.
+    const stop = lines.findIndex((line) => /^#{1,3}\s*VERDICT/i.test(line))
+    return (stop === -1 ? lines : lines.slice(0, stop)).join("\n").trim()
+  }
+  let end = lines.length
+  for (let i = start + 1; i < lines.length; i += 1) {
+    if (/^#{1,3}\s*[A-Z]/i.test(lines[i])) {
+      end = i
+      break
+    }
+  }
+  return lines.slice(start + 1, end).join("\n").trim()
+}
+
+/**
+ * One seat's tally line, plus its findings when this is the final round (#670).
+ *
+ * Only the final round expands: printing every round's full report would bury the
+ * conclusion under five historical copies, and what the caller needs is "what is
+ * still wrong", not a transcript.
+ */
+function verdictLines(v: ExpertVerdict, expand: boolean): string[] {
+  const model = v.resolvedModel ? `${v.requestedModel} -> ${v.resolvedModel}` : v.requestedModel
+  const note = v.error ? ` (${v.error})` : ""
+  const lines = [`- [${v.role}] ${v.id} (${model}): ${v.verdict} confidence=${v.confidence}${note}`]
+  if (!expand) return lines
+  const excerpt = findingsExcerpt(v.report)
+  if (excerpt === "") return lines
+  for (const line of excerpt.split(/\r?\n/)) lines.push(`    ${line}`)
+  return lines
+}
+
 export function formatOutcome(outcome: ReviewOutcome): string {
   const lines: string[] = []
+  const lastRound = outcome.rounds.length
   for (const r of outcome.rounds) {
     lines.push(`## Round ${r.round} — approvals ${r.approvals}/${r.validCount} valid`)
-    for (const v of r.verdicts) {
-      const model = v.resolvedModel ? `${v.requestedModel} -> ${v.resolvedModel}` : v.requestedModel
-      const note = v.error ? ` (${v.error})` : ""
-      lines.push(`- [${v.role}] ${v.id} (${model}): ${v.verdict} confidence=${v.confidence}${note}`)
-    }
+    const expand = r.round === lastRound
+    for (const v of r.verdicts) lines.push(...verdictLines(v, expand))
     lines.push("")
   }
   lines.push(

package/plugins/opencode/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@boyingliu01/opencode-plugin",
-  "version": "0.35.3",
+  "version": "0.35.5",
   "type": "module",
   "main": "index.ts",
   "description": "XP-Gate quality gates + AI workflow skills + Sprint Flow TUI sidebar for OpenCode",

package/plugins/opencode/skills/delphi-review/AGENTS.md

@@ -1,9 +1,9 @@
 # SKILLS/DELPHI-REVIEW KNOWLEDGE BASE
 
 **Generated:** 2026-10-30
-**Commit:** 9ebd1048
+**Commit:** 4776dc7f
 **Branch:** main
-**Version:** 0.35.3.0
+**Version:** 0.35.5.0
 
 ## OVERVIEW
 Delphi Consensus Review — multi-round anonymous expert review (≥90% threshold, exactly 3 experts with distinct executable model IDs). Supports design, requirements, and code-walkthrough modes.

package/plugins/opencode/skills/sprint-flow/AGENTS.md

@@ -1,9 +1,9 @@
 # SKILLS/SPRINT-FLOW KNOWLEDGE BASE
 
 **Generated:** 2026-10-30
-**Commit:** 9ebd1048
+**Commit:** 4776dc7f
 **Branch:** main
-**Version:** 0.35.3.0
+**Version:** 0.35.5.0
 
 ## OVERVIEW
 **6-phase** development pipeline (v2.0 compact redesign, Issue #290): PREP → DESIGN → BUILD → VERIFY → SHIP → CLOSE. Phase 3/6 BUILD default build mode is **ralph-loop** (REQ-level iteration, 40-67% token savings vs parallel). HARD-GATE between DESIGN (2/6) and BUILD (3/6): design must pass Delphi review (≥90% consensus) before any coding.

package/plugins/opencode/skills/test-specification-alignment/AGENTS.md

@@ -1,9 +1,9 @@
 # SKILLS/TEST-SPECIFICATION-ALIGNMENT KNOWLEDGE BASE
 
 **Generated:** 2026-10-30
-**Commit:** 9ebd1048
+**Commit:** 4776dc7f
 **Branch:** main
-**Version:** 0.35.3.0
+**Version:** 0.35.5.0
 
 ## OVERVIEW
 Test-Specification Alignment Engine — two-stage validation ensuring tests accurately reflect requirements and design specs.

package/plugins/qoder/plugin.json

@@ -1,6 +1,6 @@
 {
   "name": "xp-gate",
-  "version": "0.35.3",
+  "version": "0.35.5",
   "displayName": "XP-Gate",
   "description": "Extreme Programming quality gates + AI workflow skills for Qoder. Includes 12 pre-commit gates (Gate 0-11), 8 pre-push gates, Sprint Flow (6 phases), and Delphi multi-expert review (>=90% consensus).",
   "author": {

package/plugins/qoder/skills/delphi-review/AGENTS.md

@@ -1,9 +1,9 @@
 # SKILLS/DELPHI-REVIEW KNOWLEDGE BASE
 
 **Generated:** 2026-10-30
-**Commit:** 9ebd1048
+**Commit:** 4776dc7f
 **Branch:** main
-**Version:** 0.35.3.0
+**Version:** 0.35.5.0
 
 ## OVERVIEW
 Delphi Consensus Review — multi-round anonymous expert review (≥90% threshold, exactly 3 experts with distinct executable model IDs). Supports design, requirements, and code-walkthrough modes.

package/plugins/qoder/skills/sprint-flow/AGENTS.md

@@ -1,9 +1,9 @@
 # SKILLS/SPRINT-FLOW KNOWLEDGE BASE
 
 **Generated:** 2026-10-30
-**Commit:** 9ebd1048
+**Commit:** 4776dc7f
 **Branch:** main
-**Version:** 0.35.3.0
+**Version:** 0.35.5.0
 
 ## OVERVIEW
 **6-phase** development pipeline (v2.0 compact redesign, Issue #290): PREP → DESIGN → BUILD → VERIFY → SHIP → CLOSE. Phase 3/6 BUILD default build mode is **ralph-loop** (REQ-level iteration, 40-67% token savings vs parallel). HARD-GATE between DESIGN (2/6) and BUILD (3/6): design must pass Delphi review (≥90% consensus) before any coding.

package/plugins/qoder/skills/test-specification-alignment/AGENTS.md

@@ -1,9 +1,9 @@
 # SKILLS/TEST-SPECIFICATION-ALIGNMENT KNOWLEDGE BASE
 
 **Generated:** 2026-10-30
-**Commit:** 9ebd1048
+**Commit:** 4776dc7f
 **Branch:** main
-**Version:** 0.35.3.0
+**Version:** 0.35.5.0
 
 ## OVERVIEW
 Test-Specification Alignment Engine — two-stage validation ensuring tests accurately reflect requirements and design specs.

package/principles/AGENTS.md

@@ -1,9 +1,9 @@
 # PRINCIPLES CHECKER MODULE
 
 **Generated:** 2026-10-30
-**Commit:** 9ebd1048
+**Commit:** 4776dc7f
 **Branch:** main
-**Version:** 0.35.3.0
+**Version:** 0.35.5.0
 
 ## OVERVIEW
 Clean Code & SOLID principles checker — **Gate 4** of pre-commit. 14 rules × 9 language adapters, SARIF 2.1.0 output. Houses the **Boy Scout Rule** enforcement engine (Gate 6) and warning-baseline storage.

package/skills/delphi-review/AGENTS.md

@@ -1,9 +1,9 @@
 # SKILLS/DELPHI-REVIEW KNOWLEDGE BASE
 
 **Generated:** 2026-10-30
-**Commit:** 9ebd1048
+**Commit:** 4776dc7f
 **Branch:** main
-**Version:** 0.35.3.0
+**Version:** 0.35.5.0
 
 ## OVERVIEW
 Delphi Consensus Review — multi-round anonymous expert review (≥90% threshold, exactly 3 experts with distinct executable model IDs). Supports design, requirements, and code-walkthrough modes.

package/skills/sprint-flow/AGENTS.md

@@ -1,9 +1,9 @@
 # SKILLS/SPRINT-FLOW KNOWLEDGE BASE
 
 **Generated:** 2026-10-30
-**Commit:** 9ebd1048
+**Commit:** 4776dc7f
 **Branch:** main
-**Version:** 0.35.3.0
+**Version:** 0.35.5.0
 
 ## OVERVIEW
 **6-phase** development pipeline (v2.0 compact redesign, Issue #290): PREP → DESIGN → BUILD → VERIFY → SHIP → CLOSE. Phase 3/6 BUILD default build mode is **ralph-loop** (REQ-level iteration, 40-67% token savings vs parallel). HARD-GATE between DESIGN (2/6) and BUILD (3/6): design must pass Delphi review (≥90% consensus) before any coding.

package/skills/test-specification-alignment/AGENTS.md

@@ -1,9 +1,9 @@
 # SKILLS/TEST-SPECIFICATION-ALIGNMENT KNOWLEDGE BASE
 
 **Generated:** 2026-10-30
-**Commit:** 9ebd1048
+**Commit:** 4776dc7f
 **Branch:** main
-**Version:** 0.35.3.0
+**Version:** 0.35.5.0
 
 ## OVERVIEW
 Test-Specification Alignment Engine — two-stage validation ensuring tests accurately reflect requirements and design specs.