ata-coder 2.4.8__tar.gz → 2.4.9__tar.gz

{ata_coder-2.4.8/ata_coder.egg-info → ata_coder-2.4.9}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: ata-coder
-Version: 2.4.8
+Version: 2.4.9
 Summary: ATA Coder — AI-powered coding assistant
 Author: ATA Coder Team
 License-Expression: MIT
@@ -8,7 +8,7 @@ Requires-Python: >=3.10
 Description-Content-Type: text/markdown
 License-File: LICENSE
 Requires-Dist: click>=8.0
-Requires-Dist: httpx>=0.27.0
+Requires-Dist: httpx<1.0,>=0.27.0
 Requires-Dist: colorama>=0.4.6
 Requires-Dist: python-dotenv>=1.0.0
 Requires-Dist: rich>=13.0.0
@@ -21,13 +21,15 @@ Requires-Dist: pytest-timeout>=2.0; extra == "dev"
 Requires-Dist: tiktoken>=0.5.0; extra == "dev"
 Dynamic: license-file
 
-# ATA Coder v2.4.7
+# ATA Coder v2.4.8
 
 **AI-powered coding assistant — async, AST-aware, single config file.**
 
 [English](#english) | [中文](#中文)
 
-> **v2.4.7** — ⚡ **Context Memory Refactor**: O(1) token tracking, ContextManager, section-level prompt caching, pre-tokenized TF-IDF, LRU token cache. ~60% less overhead in the hot loop.
+> **v2.4.8** — 🤖 **Self-Bootstrapped Audit (Round 12)**: ATA Coder found 10 bugs in its own source — thread races, command injection, silent corruption. All self-found, all self-fixed.
+>
+> > **v2.4.7** — ⚡ **Context Memory Refactor**: O(1) token tracking, ContextManager, section-level prompt caching, pre-tokenized TF-IDF, LRU token cache. ~60% less overhead in the hot loop.
 >
 > > **v2.4.6** — 🔐 **OS-Native Credential Store**: API key encrypted at rest via Windows DPAPI / macOS Keychain / Linux secret-tool. Auto-migrates plaintext keys. Zero dependencies.
 >
@@ -552,7 +554,9 @@ All 6 findings in this release were discovered by **ATA Coder scanning its own s
 
 ## 中文
 
-> **v2.4.7** — ⚡ **上下文记忆重构**: O(1) Token 追踪、ContextManager、章节级提示缓存、预分词 TF-IDF、LRU Token 缓存。热路径开销降低约 60%。
+> **v2.4.8** — 🤖 **自举审计（第 12 轮）**: ATA Coder 审计自身源码发现 10 个 bug — 线程竞态、命令注入、静默数据损坏。全部自发现、自修复。
+>
+> > **v2.4.7** — ⚡ **上下文记忆重构**: O(1) Token 追踪、ContextManager、章节级提示缓存、预分词 TF-IDF、LRU Token 缓存。热路径开销降低约 60%。
 >
 > > **v2.4.6** — 🔐 **操作系统凭据存储**: API Key 通过 Windows DPAPI / macOS Keychain / Linux secret-tool 加密存储。自动迁移明文密钥。零依赖。
 >

{ata_coder-2.4.8 → ata_coder-2.4.9}/README.md

@@ -1,10 +1,12 @@
-# ATA Coder v2.4.7
+# ATA Coder v2.4.8
 
 **AI-powered coding assistant — async, AST-aware, single config file.**
 
 [English](#english) | [中文](#中文)
 
-> **v2.4.7** — ⚡ **Context Memory Refactor**: O(1) token tracking, ContextManager, section-level prompt caching, pre-tokenized TF-IDF, LRU token cache. ~60% less overhead in the hot loop.
+> **v2.4.8** — 🤖 **Self-Bootstrapped Audit (Round 12)**: ATA Coder found 10 bugs in its own source — thread races, command injection, silent corruption. All self-found, all self-fixed.
+>
+> > **v2.4.7** — ⚡ **Context Memory Refactor**: O(1) token tracking, ContextManager, section-level prompt caching, pre-tokenized TF-IDF, LRU token cache. ~60% less overhead in the hot loop.
 >
 > > **v2.4.6** — 🔐 **OS-Native Credential Store**: API key encrypted at rest via Windows DPAPI / macOS Keychain / Linux secret-tool. Auto-migrates plaintext keys. Zero dependencies.
 >
@@ -529,7 +531,9 @@ All 6 findings in this release were discovered by **ATA Coder scanning its own s
 
 ## 中文
 
-> **v2.4.7** — ⚡ **上下文记忆重构**: O(1) Token 追踪、ContextManager、章节级提示缓存、预分词 TF-IDF、LRU Token 缓存。热路径开销降低约 60%。
+> **v2.4.8** — 🤖 **自举审计（第 12 轮）**: ATA Coder 审计自身源码发现 10 个 bug — 线程竞态、命令注入、静默数据损坏。全部自发现、自修复。
+>
+> > **v2.4.7** — ⚡ **上下文记忆重构**: O(1) Token 追踪、ContextManager、章节级提示缓存、预分词 TF-IDF、LRU Token 缓存。热路径开销降低约 60%。
 >
 > > **v2.4.6** — 🔐 **操作系统凭据存储**: API Key 通过 Windows DPAPI / macOS Keychain / Linux secret-tool 加密存储。自动迁移明文密钥。零依赖。
 >

{ata_coder-2.4.8 → ata_coder-2.4.9}/agent.py

@@ -22,7 +22,6 @@ The agent runs a conversation loop:
 import asyncio
 import json
 import logging
-import os
 import time
 from typing import Any, Callable
 

{ata_coder-2.4.8 → ata_coder-2.4.9}/anthropic_client.py

@@ -218,7 +218,6 @@ class AnthropicClient(BaseLLMClient):
         body = sanitize_surrogates(body)
 
         # Retry loop for streaming (up to 2 retries for 429/5xx)
-        last_error = None
         for attempt in range(self._max_retries):
             try:
                 resp = await self._client.send(

{ata_coder-2.4.8 → ata_coder-2.4.9/ata_coder.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: ata-coder
-Version: 2.4.8
+Version: 2.4.9
 Summary: ATA Coder — AI-powered coding assistant
 Author: ATA Coder Team
 License-Expression: MIT
@@ -8,7 +8,7 @@ Requires-Python: >=3.10
 Description-Content-Type: text/markdown
 License-File: LICENSE
 Requires-Dist: click>=8.0
-Requires-Dist: httpx>=0.27.0
+Requires-Dist: httpx<1.0,>=0.27.0
 Requires-Dist: colorama>=0.4.6
 Requires-Dist: python-dotenv>=1.0.0
 Requires-Dist: rich>=13.0.0
@@ -21,13 +21,15 @@ Requires-Dist: pytest-timeout>=2.0; extra == "dev"
 Requires-Dist: tiktoken>=0.5.0; extra == "dev"
 Dynamic: license-file
 
-# ATA Coder v2.4.7
+# ATA Coder v2.4.8
 
 **AI-powered coding assistant — async, AST-aware, single config file.**
 
 [English](#english) | [中文](#中文)
 
-> **v2.4.7** — ⚡ **Context Memory Refactor**: O(1) token tracking, ContextManager, section-level prompt caching, pre-tokenized TF-IDF, LRU token cache. ~60% less overhead in the hot loop.
+> **v2.4.8** — 🤖 **Self-Bootstrapped Audit (Round 12)**: ATA Coder found 10 bugs in its own source — thread races, command injection, silent corruption. All self-found, all self-fixed.
+>
+> > **v2.4.7** — ⚡ **Context Memory Refactor**: O(1) token tracking, ContextManager, section-level prompt caching, pre-tokenized TF-IDF, LRU token cache. ~60% less overhead in the hot loop.
 >
 > > **v2.4.6** — 🔐 **OS-Native Credential Store**: API key encrypted at rest via Windows DPAPI / macOS Keychain / Linux secret-tool. Auto-migrates plaintext keys. Zero dependencies.
 >
@@ -552,7 +554,9 @@ All 6 findings in this release were discovered by **ATA Coder scanning its own s
 
 ## 中文
 
-> **v2.4.7** — ⚡ **上下文记忆重构**: O(1) Token 追踪、ContextManager、章节级提示缓存、预分词 TF-IDF、LRU Token 缓存。热路径开销降低约 60%。
+> **v2.4.8** — 🤖 **自举审计（第 12 轮）**: ATA Coder 审计自身源码发现 10 个 bug — 线程竞态、命令注入、静默数据损坏。全部自发现、自修复。
+>
+> > **v2.4.7** — ⚡ **上下文记忆重构**: O(1) Token 追踪、ContextManager、章节级提示缓存、预分词 TF-IDF、LRU Token 缓存。热路径开销降低约 60%。
 >
 > > **v2.4.6** — 🔐 **操作系统凭据存储**: API Key 通过 Windows DPAPI / macOS Keychain / Linux secret-tool 加密存储。自动迁移明文密钥。零依赖。
 >

{ata_coder-2.4.8 → ata_coder-2.4.9}/ata_coder.egg-info/SOURCES.txt

@@ -38,6 +38,7 @@ repl_ui.py
 safety_guard.py
 self_correct.py
 server.py
+server_rate_limit.py
 server_session.py
 server_shell.py
 session.py
@@ -89,6 +90,7 @@ utils.py
 ./safety_guard.py
 ./self_correct.py
 ./server.py
+./server_rate_limit.py
 ./server_session.py
 ./server_shell.py
 ./session.py
@@ -154,6 +156,7 @@ utils.py
 ./tools/__init__.py
 ./tools/definitions.py
 ./tools/executor.py
+./tools/file_ops.py
 ./tools/result.py
 ./tools/strategy.py
 ./tools/subagent.py
@@ -238,6 +241,7 @@ tests/test_tools.py
 tools/__init__.py
 tools/definitions.py
 tools/executor.py
+tools/file_ops.py
 tools/result.py
 tools/strategy.py
 tools/subagent.py

{ata_coder-2.4.8 → ata_coder-2.4.9}/ata_coder.egg-info/requires.txt

@@ -1,5 +1,5 @@
 click>=8.0
-httpx>=0.27.0
+httpx<1.0,>=0.27.0
 colorama>=0.4.6
 python-dotenv>=1.0.0
 rich>=13.0.0

{ata_coder-2.4.8 → ata_coder-2.4.9}/change_tracker.py

@@ -173,7 +173,7 @@ class ChangeTracker:
         if old_content == new_content:
             return None
 
-        path = Path(file_path)
+        Path(file_path)
         # Backup before edit (for undo)
         self._backup(file_path)
 

{ata_coder-2.4.8 → ata_coder-2.4.9}/llm_client.py

@@ -282,7 +282,6 @@ class LLMClient(BaseLLMClient):
             body.pop("temperature", None)
 
         # Retry loop for streaming (up to 2 retries for 429/5xx)
-        last_error = None
 
         # Sanitize surrogates before JSON encoding (prevent UTF-8 encode crash)
         from .utils import sanitize_surrogates

{ata_coder-2.4.8 → ata_coder-2.4.9}/main.py

@@ -44,7 +44,7 @@ if sys.platform == 'win32':
         _patched_init.__ata_patched__ = True
         _sp.Popen.__init__ = _patched_init
 
-__version__ = "2.4.8"
+__version__ = "2.4.9"
 
 import asyncio
 import logging

{ata_coder-2.4.8 → ata_coder-2.4.9}/prompt_template.py

@@ -240,7 +240,7 @@ class PromptTemplate:
         import re as _re
 
         if_tag = _re.compile(r'\{\%\s*if\s+(.+?)\s*\%\}')
-        endif_tag = _re.compile(r'\{\%\s*endif\s*\%\}')
+        _re.compile(r'\{\%\s*endif\s*\%\}')
         any_tag = _re.compile(r'\{\%\s*(?:if\s+.+?|endif)\s*\%\}')
 
         def _eval_condition(cond: str) -> bool:

{ata_coder-2.4.8 → ata_coder-2.4.9}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "ata-coder"
-version = "2.4.8"
+version = "2.4.9"
 description = "ATA Coder — AI-powered coding assistant"
 readme = "README.md"
 requires-python = ">=3.10"
@@ -14,7 +14,7 @@ authors = [
 ]
 dependencies = [
     "click>=8.0",
-    "httpx>=0.27.0",
+    "httpx>=0.27.0,<1.0",
     "colorama>=0.4.6",
     "python-dotenv>=1.0.0",
     "rich>=13.0.0",
@@ -81,16 +81,22 @@ exclude = ["build", "dist", ".git", "__pycache__", ".pytest_cache", "web/node_mo
 
 [tool.ruff.lint]
 select = ["E", "F", "W"]
-ignore = ["E501", "E402", "E701", "E722", "E741", "E731", "F401", "F821", "F841"]
+ignore = ["E501", "E402", "E701", "E722", "E741", "E731", "F821"]
+
+[tool.ruff.lint.per-file-ignores]
+# repl_ui.py + terminal.py: rich/colorama imports are availability checks
+"repl_ui.py" = ["F401"]
+"terminal.py" = ["F401"]
+# Skill handlers + tests: some imports are for dynamic loading or optional deps
+"skills/*/handler.py" = ["F401"]
+"tests/*.py" = ["F401", "F841"]
 # E501: line too long (enforced by formatter instead)
 # E402: imports after sys.path.insert (intentional in main/server/gui)
 # E701: multiple statements on one line (compact error handling)
 # E722: bare except (intentional for cleanup/shutdown paths)
 # E741: ambiguous variable name (l for loop vars)
 # E731: lambda assignment (trivial inline helpers, not exported)
-# F401: unused import (optional deps guarded by try/except)
 # F821: undefined name (optional deps guarded by try/except)
-# F841: unused variable (test assertions via side effects)
 
 [tool.mypy]
 python_version = "3.10"

{ata_coder-2.4.8 → ata_coder-2.4.9}/server.py

@@ -22,8 +22,6 @@ Usage:
   python main.py --server                 # From main launcher
 """
 
-import asyncio
-import collections
 import json
 import logging
 import os
@@ -101,6 +99,7 @@ from .config import AppConfig, get_config
 from .tools import TOOL_DEFINITIONS
 from .server_session import SessionStore
 from .server_shell import shell_open, shell_ensure, shell_close, shell_close_all, get_shell_sessions
+from .server_rate_limit import RateLimiter
 from .skills import get_skill_manager
 from .utils import brief_args
 
@@ -112,11 +111,13 @@ logger = logging.getLogger(__name__)
 # ══════════════════════════════════════════════════════════════════════# HTTP Request Handler
 # ═══════════════════════════════════════════════════════════════════════════════
 
-class AgentAPIHandler(BaseHTTPRequestHandler):
+class AgentAPIHandler(RateLimiter, BaseHTTPRequestHandler):
     """HTTP handler for the ATA Coder API.
 
     *config* and *store* are set as class attributes by :func:`create_server`
     before the server starts accepting requests.
+
+    Rate limiting is inherited from :class:`RateLimiter`.
     """
 
     # Class-level references (set by server factory before accepting requests).
@@ -126,80 +127,6 @@ class AgentAPIHandler(BaseHTTPRequestHandler):
     store: "SessionStore | None" = None
     _ws_lock: threading.Lock = threading.Lock()  # protects workspace dir reads/writes
 
-    # ── Rate limiting (class-level, shared across handler instances) ──────
-    _rate_lock: threading.Lock = threading.Lock()
-    _rate_buckets: dict[str, "collections.deque[float]"] = {}  # ip → deque of timestamps
-    _rate_blocked: dict[str, float] = {}  # ip → block expiry timestamp
-    _RATE_MAX_REQUESTS = 120   # max requests per window
-    _RATE_WINDOW_S = 60.0      # sliding window in seconds
-    _RATE_BLOCK_S = 300.0      # block duration after exceeding penalty threshold
-    _RATE_PENALTY_MULTIPLIER = 3  # requests × this = block threshold
-    _RATE_CLEANUP_INTERVAL = 1000  # amortized: trigger cleanup every N calls
-    _rate_cleanup_counter: int = 0
-
-    @classmethod
-    def _cleanup_rate_buckets(cls, now: float) -> None:
-        """Remove stale IP entries whose last activity exceeds 2× the window.
-
-        Without this, IPs that stay under the rate limit forever accumulate
-        in _rate_buckets and leak memory over long-running server processes.
-        """
-        cutoff = now - cls._RATE_WINDOW_S * 2
-        stale = [
-            ip for ip, dq in cls._rate_buckets.items()
-            if not dq or dq[-1] <= cutoff
-        ]
-        for ip in stale:
-            del cls._rate_buckets[ip]
-        if stale:
-            logger.debug("Rate limiter: pruned %d stale IP bucket(s)", len(stale))
-
-    @classmethod
-    def _check_rate_limit(cls, client_ip: str) -> bool:
-        """Sliding-window rate limiter with deque for O(1) cleanup.
-
-        Returns True if request is allowed.
-        """
-        now = time.time()
-        with cls._rate_lock:
-            # Periodic stale-bucket cleanup (amortized — triggered every N calls)
-            cls._rate_cleanup_counter += 1
-            if cls._rate_cleanup_counter >= cls._RATE_CLEANUP_INTERVAL:
-                cls._rate_cleanup_counter = 0
-                cls._cleanup_rate_buckets(now)
-
-            # Check if IP is currently blocked (penalty tier)
-            blocked_until = cls._rate_blocked.get(client_ip, 0)
-            if now < blocked_until:
-                return False
-            if now >= blocked_until and client_ip in cls._rate_blocked:
-                del cls._rate_blocked[client_ip]
-
-            dq = cls._rate_buckets.get(client_ip)
-            if dq is None:
-                dq = collections.deque()
-                cls._rate_buckets[client_ip] = dq
-
-            # Purge expired entries — O(1) per entry via popleft
-            cutoff = now - cls._RATE_WINDOW_S
-            while dq and dq[0] <= cutoff:
-                dq.popleft()
-
-            # Penalty tier: block if request count exceeds penalty threshold
-            penalty_limit = cls._RATE_MAX_REQUESTS * cls._RATE_PENALTY_MULTIPLIER
-            if len(dq) > penalty_limit:
-                cls._rate_blocked[client_ip] = now + cls._RATE_BLOCK_S
-                logger.warning("Rate limit BLOCK: %s for %ds (%d requests in window)",
-                               client_ip, cls._RATE_BLOCK_S, len(dq))
-                return False
-
-            # Standard rate limit
-            if len(dq) >= cls._RATE_MAX_REQUESTS:
-                return False
-
-            dq.append(now)
-        return True
-
     def __init__(self, *args, **kwargs):
         # Per-instance copies for thread-safe access under ThreadingHTTPServer
         self.config = self.__class__.config
@@ -469,7 +396,6 @@ class AgentAPIHandler(BaseHTTPRequestHandler):
         except Exception:
             logger.debug("Failed to fetch models from API, using cache", exc_info=True)
         # Fallback: cached model list from settings or env
-        import os
         from .settings import get_settings
         cached = get_settings().get("env", "ATA_CODER_MODELS_CACHE", default="") or self.config.llm.model
         models = [{"id": m.strip(), "owned_by": ""} for m in cached.split(",") if m.strip()]

ata_coder-2.4.9/server_rate_limit.py

@@ -0,0 +1,89 @@
+"""Sliding-window rate limiter for the HTTP API server.
+
+Extracted from server.py to reduce file size and isolate concerns.
+Used by AgentAPIHandler as a class-level mixin.
+"""
+
+import collections
+import logging
+import threading
+import time
+
+logger = logging.getLogger(__name__)
+
+
+class RateLimiter:
+    """Sliding-window rate limiter with deque for O(1) expiry + penalty tier.
+
+    Usage (as class-level mixin on a BaseHTTPRequestHandler subclass):
+        class MyHandler(RateLimiter, BaseHTTPRequestHandler):
+            ...
+            allowed = self._check_rate_limit(self.client_address[0])
+    """
+
+    # ── Configuration ──────────────────────────────────────────────────────
+    _rate_lock: threading.Lock = threading.Lock()
+    _rate_buckets: dict[str, "collections.deque[float]"] = {}  # ip → deque of timestamps
+    _rate_blocked: dict[str, float] = {}  # ip → block expiry timestamp
+    _RATE_MAX_REQUESTS = 120   # max requests per window
+    _RATE_WINDOW_S = 60.0      # sliding window in seconds
+    _RATE_BLOCK_S = 300.0      # block duration after exceeding penalty threshold
+    _RATE_PENALTY_MULTIPLIER = 3  # requests × this = block threshold
+    _RATE_CLEANUP_INTERVAL = 1000  # amortized: trigger cleanup every N calls
+    _rate_cleanup_counter: int = 0
+
+    @classmethod
+    def _cleanup_rate_buckets(cls, now: float) -> None:
+        """Remove stale IP entries whose last activity exceeds 2× the window."""
+        cutoff = now - cls._RATE_WINDOW_S * 2
+        stale = [
+            ip for ip, dq in cls._rate_buckets.items()
+            if not dq or dq[-1] <= cutoff
+        ]
+        for ip in stale:
+            del cls._rate_buckets[ip]
+        if stale:
+            logger.debug("Rate limiter: pruned %d stale IP bucket(s)", len(stale))
+
+    @classmethod
+    def _check_rate_limit(cls, client_ip: str) -> bool:
+        """Sliding-window rate limiter. Returns True if request is allowed."""
+        now = time.time()
+        with cls._rate_lock:
+            # Periodic stale-bucket cleanup (amortized)
+            cls._rate_cleanup_counter += 1
+            if cls._rate_cleanup_counter >= cls._RATE_CLEANUP_INTERVAL:
+                cls._rate_cleanup_counter = 0
+                cls._cleanup_rate_buckets(now)
+
+            # Check if IP is currently blocked (penalty tier)
+            blocked_until = cls._rate_blocked.get(client_ip, 0)
+            if now < blocked_until:
+                return False
+            if now >= blocked_until and client_ip in cls._rate_blocked:
+                del cls._rate_blocked[client_ip]
+
+            dq = cls._rate_buckets.get(client_ip)
+            if dq is None:
+                dq = collections.deque()
+                cls._rate_buckets[client_ip] = dq
+
+            # Purge expired entries — O(1) per entry via popleft
+            cutoff = now - cls._RATE_WINDOW_S
+            while dq and dq[0] <= cutoff:
+                dq.popleft()
+
+            # Penalty tier: block if request count exceeds penalty threshold
+            penalty_limit = cls._RATE_MAX_REQUESTS * cls._RATE_PENALTY_MULTIPLIER
+            if len(dq) > penalty_limit:
+                cls._rate_blocked[client_ip] = now + cls._RATE_BLOCK_S
+                logger.warning("Rate limit BLOCK: %s for %ds (%d requests in window)",
+                               client_ip, cls._RATE_BLOCK_S, len(dq))
+                return False
+
+            # Standard rate limit
+            if len(dq) >= cls._RATE_MAX_REQUESTS:
+                return False
+
+            dq.append(now)
+        return True

{ata_coder-2.4.8 → ata_coder-2.4.9}/settings.py

@@ -541,7 +541,6 @@ def _store_credential(service: str, account: str, secret: str) -> bool:
                 cred_file.write_text(result.stdout.strip(), encoding="utf-8")
                 # Restrictive permissions on the credential file
                 try:
-                    import stat
                     cred_file.chmod(0o600)
                 except Exception:
                     pass

{ata_coder-2.4.8 → ata_coder-2.4.9}/setup_wizard.py

@@ -111,7 +111,7 @@ def run_setup_wizard() -> None:
     print()
 
 
-__version__ = "2.4.8"
+__version__ = "2.4.9"
 
 
 def print_banner() -> None:

{ata_coder-2.4.8 → ata_coder-2.4.9}/skills/math-calculator/handler.py

@@ -4,7 +4,6 @@
 import ast
 import math
 import operator
-import re
 from typing import Any
 
 
@@ -155,7 +154,7 @@ def run(input_data: dict[str, Any]) -> dict[str, Any]:
     """
     expression = input_data.get("expression", "").strip()
     precision = input_data.get("precision", 6)
-    output_format = input_data.get("format", "number")
+    input_data.get("format", "number")
 
     # Validate
     if not expression:

{ata_coder-2.4.8 → ata_coder-2.4.9}/sub_agent.py

@@ -12,7 +12,6 @@ Each SubAgent runs as an asyncio.Task and has:
 import asyncio
 import json
 import logging
-import os
 import uuid
 from dataclasses import dataclass, field
 from typing import Callable, Optional

{ata_coder-2.4.8 → ata_coder-2.4.9}/tests/test_change_tracker.py

@@ -110,7 +110,7 @@ class TestChangeTracker:
 
     def test_init_creates_backup_dir(self, tmp_path):
         """ChangeTracker should create backup directory."""
-        tracker = ChangeTracker("test-session", backup_dir=tmp_path / "backups")
+        ChangeTracker("test-session", backup_dir=tmp_path / "backups")
         assert (tmp_path / "backups").exists()
 
     def test_capture_write_new_file(self, tmp_path):

{ata_coder-2.4.8 → ata_coder-2.4.9}/tests/test_memory.py

@@ -144,13 +144,13 @@ class TestMemoryStoreInit:
     def test_init_creates_memory_dir(self, tmp_path):
         """MemoryStore should create the memory directory on init."""
         memory_dir = tmp_path / "ata_memories"
-        store = MemoryStore(memory_dir)
+        MemoryStore(memory_dir)
         assert memory_dir.exists()
 
     def test_init_creates_memory_index(self, tmp_path):
         """MemoryStore should create MEMORY.md index on init."""
         memory_dir = tmp_path / "ata_memories"
-        store = MemoryStore(memory_dir)
+        MemoryStore(memory_dir)
         index_file = memory_dir / "MEMORY.md"
         assert index_file.exists()
 

{ata_coder-2.4.8 → ata_coder-2.4.9}/tests/test_tools.py

@@ -147,7 +147,7 @@ class TestToolReadFile:
     async def test_read_caches_file(self):
         f = Path(self.tmp) / "cache_test.txt"
         f.write_text("cached content\n", encoding="utf-8")
-        cache_key = str(f.resolve())
+        str(f.resolve())
         # First read
         r1 = await self.executor.execute("read_file", {"file_path": str(f)})
         assert r1.success