synapse-hub 2.2.4 → 2.4.0

package/tools/git_pre_commit_guard.py

@@ -11,6 +11,7 @@ Ensures:
 
 import sys
 import os
+import shlex
 import subprocess
 
 # Enforce UTF-8 stdout encoding for Windows console compatibility
@@ -23,8 +24,13 @@ if hasattr(sys.stdout, 'reconfigure'):
 
 FORBIDDEN_PATHS = ["/Users/", "C:\\Workspace\\", "D:\\Workspace\\", "file:///c:", "file:///d:"]
 
-def run_cmd(cmd):
-    return subprocess.run(cmd, stdout=subprocess.PIPE, stderr=subprocess.PIPE, text=True, shell=True)
+def run_cmd(cmd, timeout: float = 60.0):
+    if isinstance(cmd, str):
+        cmd = shlex.split(cmd)
+    try:
+        return subprocess.run(cmd, stdout=subprocess.PIPE, stderr=subprocess.PIPE, text=True, timeout=timeout)
+    except subprocess.TimeoutExpired:
+        return subprocess.CompletedProcess(cmd, returncode=124, stdout="", stderr=f"Command timed out after {timeout}s")
 
 def check_bom(files):
     bom_files = []
@@ -187,7 +193,7 @@ def main():
     if doc_affected:
         linter_script = os.path.join(os.path.dirname(__file__), "lint_ai_docs.py")
         if os.path.exists(linter_script):
-            res = subprocess.run([sys.executable, linter_script], stdout=subprocess.PIPE, stderr=subprocess.PIPE, text=True)
+            res = run_cmd([sys.executable, linter_script], timeout=60.0)
             if res.returncode != 0:
                 print("❌ [PRE-COMMIT BLOCKED] AI Documentation Linter detected drift:")
                 print(res.stdout.strip())
@@ -198,16 +204,16 @@ def main():
     if keywords_affected:
         repo_map_script = os.path.join(os.path.dirname(__file__), "repo_map.py")
         if os.path.exists(repo_map_script):
-            subprocess.run([sys.executable, repo_map_script], stdout=subprocess.DEVNULL, stderr=subprocess.DEVNULL)
+            run_cmd([sys.executable, repo_map_script], timeout=60.0)
             # Stage updated repo_map files
-            subprocess.run(["git", "add", "tools/repo_map.txt", "tools/repo_map.json"], stdout=subprocess.DEVNULL, stderr=subprocess.DEVNULL)
+            run_cmd(["git", "add", "tools/repo_map.txt", "tools/repo_map.json"], timeout=30.0)
 
     # 6. Scripts/ change -> Static Defect Linter (Meta Infer-Style)
     scripts_affected = any(f.startswith("Scripts/") and f.endswith(".groovy") for f in staged_files)
     if scripts_affected:
         infer_script = os.path.join(os.path.dirname(__file__), "infer_linter.py")
         if os.path.exists(infer_script):
-            res = subprocess.run([sys.executable, infer_script], stdout=subprocess.PIPE, stderr=subprocess.PIPE, text=True)
+            res = run_cmd([sys.executable, infer_script], timeout=60.0)
             if res.returncode != 0:
                 print("❌ [PRE-COMMIT BLOCKED] Static Defect Sentinel detected critical issues in staged scripts:")
                 print(res.stdout.strip())

package/tools/goldfish_review.py

@@ -0,0 +1,366 @@
+#!/usr/bin/env python3
+# SPDX-FileCopyrightText: Copyright (c) 2026 Thế Vị (@TeeVee998). All rights reserved.
+# SPDX-License-Identifier: MIT
+# Project: Synapse-Hub (https://github.com/TeeVee998/Synapse-Hub)
+
+"""
+tools/goldfish_review.py — Zero-Context Test Specification Reviewer
+-------------------------------------------------------------------
+Inspired by Google's Elephant-Goldfish Model (EGM).
+Acts as a stateless, zero-context 'Goldfish' reviewer to stress-test test specifications,
+BDD Gherkin feature files, and Acceptance Criteria BEFORE code synthesis begins.
+
+5 Verification Vectors:
+  1. Comprehension & System Boundaries (20%): Actor, Preconditions, Clear Target Scope.
+  2. Dynamic Data & Idempotent Teardown (20%): Dynamic IDs (uniqueCode), No hardcoded collisions, Teardown hook.
+  3. Assertion Rigor (25%): Unambiguous expected results, Anti-Fake Pass, explicit state checks.
+  4. Boundary & Negative Scenarios (20%): Negative flows, invalid inputs, edge cases covered.
+  5. Implementation Target (15%): Concrete Page Class mapping, API endpoints, or UI targets specified.
+
+Usage:
+  python tools/goldfish_review.py <path_to_spec.md_or_.feature>
+  python tools/goldfish_review.py --json <path_to_spec>
+"""
+
+import os
+import sys
+import re
+import json
+import argparse
+
+if hasattr(sys.stdout, 'reconfigure'):
+    try:
+        sys.stdout.reconfigure(encoding='utf-8')
+        sys.stderr.reconfigure(encoding='utf-8')
+    except Exception:
+        pass
+
+
+VAGUE_ASSERTION_PATTERNS = [
+    re.compile(r'\b(works? fine|works? properly|works? as expected)\b', re.IGNORECASE),
+    re.compile(r'\b(system works|page opens|everything is ok|no errors?)\b', re.IGNORECASE),
+    re.compile(r'\b(hệ thống hoạt động bình thường|thành công mà không có lỗi|hiển thị đúng)\b', re.IGNORECASE),
+    re.compile(r'\b(should succeed|should pass)\b(?!\s+(with|and assert|expecting))', re.IGNORECASE),
+]
+
+PRECISE_ASSERTION_PATTERNS = [
+    re.compile(r'\b(assert|expect|verify|check)\b', re.IGNORECASE),
+    re.compile(r'\b(status code|http \d{3}|returns? \d{3})\b', re.IGNORECASE),
+    re.compile(r'\b(contains?|equals?|matches?|visible|hidden|disabled|enabled)\b', re.IGNORECASE),
+    re.compile(r'\b(then\s+the\s+.*\s+(should|must|will))\b', re.IGNORECASE),
+    re.compile(r'\b(thông báo|thành công|thất bại|mã lỗi|kết quả mong đợi)\b', re.IGNORECASE),
+]
+
+DYNAMIC_DATA_PATTERNS = [
+    re.compile(r'\b(uniquecode|timestamp|random|uuid|generator|autogen|dyn_)\b', re.IGNORECASE),
+    re.compile(r'\b(auto_\w+|test_\$\w+|faker)\b', re.IGNORECASE),
+    re.compile(r'\b(dữ liệu động|ngẫu nhiên|mã tự sinh)\b', re.IGNORECASE),
+]
+
+HARDCODED_DATA_PATTERNS = [
+    re.compile(r'["\'](user\d+|admin\d+|testuser|demo\d+|pass\d+)["\']', re.IGNORECASE),
+    re.compile(r'id\s*[:=]\s*["\']\d+["\']', re.IGNORECASE),
+]
+
+TEARDOWN_PATTERNS = [
+    re.compile(r'\b(teardown|cleanup|delete after|roll back|xóa dữ liệu|dọn dẹp|aftereach|afterall)\b', re.IGNORECASE),
+]
+
+BOUNDARY_NEGATIVE_PATTERNS = [
+    re.compile(r'\b(negative|invalid|boundary|error|fail|reject|unauthorized|forbidden|400|401|403|404|500)\b', re.IGNORECASE),
+    re.compile(r'\b(trường hợp lỗi|không hợp lệ|quá độ dài|để trống|bỏ trống|sai định dạng)\b', re.IGNORECASE),
+    re.compile(r'\b(double submit|race condition|timeout|network lag)\b', re.IGNORECASE),
+]
+
+IMPLEMENTATION_TARGET_PATTERNS = [
+    re.compile(r'\b([A-Z][a-zA-Z0-9]+(Page|Screen|View|Modal|Dialog|Api|Client|Service))\b'),
+    re.compile(r'\b(page class|endpoint|url|route|table|input|button|form)\b', re.IGNORECASE),
+    re.compile(r'(/api/v\d+/[\w/]+)', re.IGNORECASE),
+]
+
+
+def evaluate_spec_content(content: str, filename: str = "") -> dict:
+    """
+    Evaluates a test specification content through the 5 Goldfish vectors.
+    Returns a structured dictionary with scores, findings, and overall readiness.
+    """
+    if not content or not content.strip():
+        return {
+            "file": filename,
+            "score": 0,
+            "passed": False,
+            "status": "REJECTED",
+            "vector_scores": {
+                "Comprehension": 0,
+                "Data Independence": 0,
+                "Assertion Rigor": 0,
+                "Boundary Coverage": 0,
+                "Implementation Target": 0
+            },
+            "findings_count": 1,
+            "findings": [{
+                "vector": "Comprehension",
+                "severity": "CRITICAL",
+                "message": "Specification content is completely empty or whitespace only."
+            }]
+        }
+
+    findings = []
+    vector_scores = {}
+
+    lines = [line.strip() for line in content.splitlines() if line.strip()]
+    full_text = content.lower()
+
+    # -------------------------------------------------------------
+    # Vector 1: Comprehension & System Boundaries (Weight: 20)
+    # -------------------------------------------------------------
+    v1_score = 0
+    has_title = bool(re.search(r'^(#|feature:|title:|user story:)', content, re.MULTILINE | re.IGNORECASE))
+    has_preconditions = bool(re.search(r'\b(precondition|given|context|tiền điều kiện|bối cảnh|trước khi)\b', content, re.IGNORECASE))
+    has_actor = bool(re.search(r'\b(as a|actor|user|admin|role|người dùng|vai trò)\b', content, re.IGNORECASE))
+
+    if has_title:
+        v1_score += 8
+    else:
+        findings.append({
+            "vector": "Comprehension",
+            "severity": "MEDIUM",
+            "message": "Missing explicit Title or Feature declaration (# Title / Feature: ...)."
+        })
+
+    if has_preconditions:
+        v1_score += 7
+    else:
+        findings.append({
+            "vector": "Comprehension",
+            "severity": "HIGH",
+            "message": "Missing Preconditions / Given setup context. Zero-context agents cannot deduce start state."
+        })
+
+    if has_actor:
+        v1_score += 5
+    else:
+        findings.append({
+            "vector": "Comprehension",
+            "severity": "LOW",
+            "message": "Missing explicit Actor/Role definition ('As a ...' / 'Actor: ...')."
+        })
+    vector_scores["Comprehension"] = v1_score
+
+    # -------------------------------------------------------------
+    # Vector 2: Dynamic Data & Idempotent Teardown (Weight: 20)
+    # -------------------------------------------------------------
+    v2_score = 0
+    has_dynamic_data = any(p.search(content) for p in DYNAMIC_DATA_PATTERNS)
+    has_hardcoded_data = any(p.search(content) for p in HARDCODED_DATA_PATTERNS)
+    has_teardown = any(p.search(content) for p in TEARDOWN_PATTERNS)
+
+    is_creation_test = bool(re.search(r'\b(create|add|insert|register|post|tạo mới|thêm mới|đăng ký)\b', content, re.IGNORECASE))
+
+    if has_dynamic_data:
+        v2_score += 10
+    elif not is_creation_test:
+        v2_score += 7  # Read-only or navigation spec
+    else:
+        findings.append({
+            "vector": "Data Independence",
+            "severity": "HIGH",
+            "message": "Creation spec lacks dynamic test data generation (uniqueCode, timestamp). High risk of DB collisions."
+        })
+
+    if has_teardown:
+        v2_score += 10
+    elif not is_creation_test:
+        v2_score += 10
+    else:
+        findings.append({
+            "vector": "Data Independence",
+            "severity": "HIGH",
+            "message": "Creation spec lacks explicit automated Teardown / Cleanup step. Violates DB hygiene discipline."
+        })
+
+    if has_hardcoded_data:
+        v2_score = max(0, v2_score - 5)
+        findings.append({
+            "vector": "Data Independence",
+            "severity": "MEDIUM",
+            "message": "Detected static/hardcoded credentials or entity IDs. Must parameterize via environment or dynamic generator."
+        })
+    vector_scores["Data Independence"] = v2_score
+
+    # -------------------------------------------------------------
+    # Vector 3: Assertion Rigor & Truth Anchor (Weight: 25)
+    # -------------------------------------------------------------
+    v3_score = 0
+    precise_matches = sum(len(p.findall(content)) for p in PRECISE_ASSERTION_PATTERNS)
+    vague_matches = sum(len(p.findall(content)) for p in VAGUE_ASSERTION_PATTERNS)
+
+    if precise_matches >= 3:
+        v3_score += 25
+    elif precise_matches >= 1:
+        v3_score += 15
+        findings.append({
+            "vector": "Assertion Rigor",
+            "severity": "MEDIUM",
+            "message": f"Low assertion density (found {precise_matches} precise check(s)). Spec needs more concrete expected values."
+        })
+    else:
+        findings.append({
+            "vector": "Assertion Rigor",
+            "severity": "CRITICAL",
+            "message": "Zero concrete assertions or expected outcomes detected in spec. Violates Truth Anchor."
+        })
+
+    if vague_matches > 0:
+        v3_score = max(0, v3_score - (vague_matches * 5))
+        findings.append({
+            "vector": "Assertion Rigor",
+            "severity": "HIGH",
+            "message": f"Found {vague_matches} vague assertion phrase(s) ('works fine', 'hoạt động bình thường'). Replace with concrete verification criteria."
+        })
+    vector_scores["Assertion Rigor"] = v3_score
+
+    # -------------------------------------------------------------
+    # Vector 4: Boundary & Negative Scenarios (Weight: 20)
+    # -------------------------------------------------------------
+    v4_score = 0
+    has_boundary = any(p.search(content) for p in BOUNDARY_NEGATIVE_PATTERNS)
+
+    if has_boundary:
+        v4_score = 20
+    else:
+        findings.append({
+            "vector": "Boundary Coverage",
+            "severity": "MEDIUM",
+            "message": "No negative, boundary, or invalid input scenarios specified. Spec only tests the happy path."
+        })
+        v4_score = 5
+    vector_scores["Boundary Coverage"] = v4_score
+
+    # -------------------------------------------------------------
+    # Vector 5: Implementation Target & Technical Mapping (Weight: 15)
+    # -------------------------------------------------------------
+    v5_score = 0
+    has_target = any(p.search(content) for p in IMPLEMENTATION_TARGET_PATTERNS)
+
+    if has_target:
+        v5_score = 15
+    else:
+        findings.append({
+            "vector": "Implementation Target",
+            "severity": "LOW",
+            "message": "No target PageClass, API endpoint, or concrete UI selector category referenced. Engineers must guess technical anchors."
+        })
+        v5_score = 5
+    vector_scores["Implementation Target"] = v5_score
+
+    # Total score calculation
+    total_score = sum(vector_scores.values())
+    total_score = max(0, min(100, total_score))
+
+    passed = total_score >= 80 and not any(f["severity"] == "CRITICAL" for f in findings)
+
+    return {
+        "file": filename,
+        "score": total_score,
+        "passed": passed,
+        "status": "APPROVED" if passed else "REJECTED",
+        "vector_scores": vector_scores,
+        "findings_count": len(findings),
+        "findings": findings
+    }
+
+
+def review_file(filepath: str) -> dict:
+    if not os.path.isfile(filepath):
+        return {
+            "file": filepath,
+            "score": 0,
+            "passed": False,
+            "status": "ERROR",
+            "vector_scores": {},
+            "findings_count": 1,
+            "findings": [{"vector": "File System", "severity": "CRITICAL", "message": f"File not found: {filepath}"}]
+        }
+
+    try:
+        with open(filepath, "r", encoding="utf-8") as f:
+            content = f.read()
+    except UnicodeDecodeError:
+        try:
+            with open(filepath, "r", encoding="latin-1") as f:
+                content = f.read()
+        except Exception as e:
+            return {
+                "file": filepath,
+                "score": 0,
+                "passed": False,
+                "status": "ERROR",
+                "vector_scores": {},
+                "findings_count": 1,
+                "findings": [{"vector": "Encoding", "severity": "CRITICAL", "message": f"Could not read file: {e}"}]
+            }
+
+    return evaluate_spec_content(content, filename=filepath)
+
+
+def print_cli_report(result: dict):
+    print("=" * 70)
+    print("  GOLDFISH SPECIFICATION AUDITOR — ZERO-CONTEXT VERIFICATION")
+    print("  Inspired by Google Elephant-Goldfish Model (EGM)")
+    print("=" * 70)
+    print(f"Target Specification : {result['file'] or 'Direct Input'}")
+    print(f"Readiness Score      : {result['score']}/100")
+    status_icon = "✅" if result['passed'] else "❌"
+    print(f"Audit Verdict        : {status_icon} [{result['status']}] (Threshold: >= 80%)")
+    print("-" * 70)
+    print("Vector Breakdown:")
+    for vec, score in result.get("vector_scores", {}).items():
+        max_v = 25 if vec == "Assertion Rigor" else (15 if vec == "Implementation Target" else 20)
+        filled = int(score / max_v * 10) if max_v > 0 else 0
+        filled = max(0, min(10, filled))
+        bar = "#" * filled + "-" * (10 - filled)
+        print(f"  • {vec:<24}: {score:>2}/{max_v:>2} [{bar}]")
+
+    print("-" * 70)
+    if result["findings"]:
+        print(f"Discovered Unstated Assumptions & Gaps ({len(result['findings'])}):")
+        for i, item in enumerate(result["findings"], 1):
+            sev = item["severity"]
+            sev_badge = f"[{sev}]"
+            print(f"  {i}. {sev_badge:<10} ({item['vector']}) {item['message']}")
+    else:
+        print("✅ Zero unstated assumptions detected. Specification is 100% implementation-ready.")
+    print("=" * 70)
+
+
+def main():
+    parser = argparse.ArgumentParser(
+        description="Goldfish Spec Reviewer — Zero-context stress-test for specifications and user stories."
+    )
+    parser.add_argument("spec_file", nargs="?", default="", help="Path to specification Markdown or Gherkin .feature file (use '-' for stdin)")
+    parser.add_argument("--stdin", action="store_true", help="Read specification from standard input (stdin)")
+    parser.add_argument("--json", action="store_true", help="Output machine-readable JSON format")
+    parser.add_argument("--min-score", type=int, default=80, help="Minimum readiness score threshold (default: 80)")
+
+    args = parser.parse_args()
+
+    if args.spec_file == "-" or getattr(args, "stdin", False):
+        content = sys.stdin.read()
+        result = evaluate_spec_content(content, filename="<stdin>")
+    elif not args.spec_file:
+        parser.print_help()
+        sys.exit(1)
+    else:
+        result = review_file(args.spec_file)
+
+    if args.json:
+        print(json.dumps(result, indent=2, ensure_ascii=False))
+    else:
+        print_cli_report(result)
+
+    sys.exit(0 if result["passed"] else 1)
+
+
+if __name__ == "__main__":
+    main()

package/tools/heal_pr_automator.py

@@ -140,7 +140,7 @@ def generate_pr_draft(
 #### 3. Verification Checklist
 - [x] Confidence score meets threshold ($\ge 90\%$)
 - [x] Tested with `python tools/test_impact.py` on impacted test cases
-- [x] 8 Quality Gates DoD satisfied
+- [x] 9 Quality Gates DoD satisfied
 {checklist_trace}{evidence_markdown}
 """
     return {