hotfix/pr-review-bug
This commit is contained in:
Test User
@@ -35,18 +35,18 @@
|
||||
> ⚠️ Beta releases may contain bugs and breaking changes. [View all releases](https://github.com/AndyMik90/Auto-Claude/releases)
|
||||
|
||||
<!-- BETA_VERSION_BADGE -->
|
||||
[](https://github.com/AndyMik90/Auto-Claude/releases/tag/v2.7.6-beta.2)
|
||||
[](https://github.com/AndyMik90/Auto-Claude/releases/tag/v2.7.2-beta.10)
|
||||
<!-- BETA_VERSION_BADGE_END -->
|
||||
|
||||
<!-- BETA_DOWNLOADS -->
|
||||
| Platform | Download |
|
||||
|----------|----------|
|
||||
| **Windows** | [Auto-Claude-2.7.6-beta.2-win32-x64.exe](https://github.com/AndyMik90/Auto-Claude/releases/download/v2.7.6-beta.2/Auto-Claude-2.7.6-beta.2-win32-x64.exe) |
|
||||
| **macOS (Apple Silicon)** | [Auto-Claude-2.7.6-beta.2-darwin-arm64.dmg](https://github.com/AndyMik90/Auto-Claude/releases/download/v2.7.6-beta.2/Auto-Claude-2.7.6-beta.2-darwin-arm64.dmg) |
|
||||
| **macOS (Intel)** | [Auto-Claude-2.7.6-beta.2-darwin-x64.dmg](https://github.com/AndyMik90/Auto-Claude/releases/download/v2.7.6-beta.2/Auto-Claude-2.7.6-beta.2-darwin-x64.dmg) |
|
||||
| **Linux** | [Auto-Claude-2.7.6-beta.2-linux-x86_64.AppImage](https://github.com/AndyMik90/Auto-Claude/releases/download/v2.7.6-beta.2/Auto-Claude-2.7.6-beta.2-linux-x86_64.AppImage) |
|
||||
| **Linux (Debian)** | [Auto-Claude-2.7.6-beta.2-linux-amd64.deb](https://github.com/AndyMik90/Auto-Claude/releases/download/v2.7.6-beta.2/Auto-Claude-2.7.6-beta.2-linux-amd64.deb) |
|
||||
| **Linux (Flatpak)** | [Auto-Claude-2.7.6-beta.2-linux-x86_64.flatpak](https://github.com/AndyMik90/Auto-Claude/releases/download/v2.7.6-beta.2/Auto-Claude-2.7.6-beta.2-linux-x86_64.flatpak) |
|
||||
| **Windows** | [Auto-Claude-2.7.2-beta.10-win32-x64.exe](https://github.com/AndyMik90/Auto-Claude/releases/download/v2.7.2-beta.10/Auto-Claude-2.7.2-beta.10-win32-x64.exe) |
|
||||
| **macOS (Apple Silicon)** | [Auto-Claude-2.7.2-beta.10-darwin-arm64.dmg](https://github.com/AndyMik90/Auto-Claude/releases/download/v2.7.2-beta.10/Auto-Claude-2.7.2-beta.10-darwin-arm64.dmg) |
|
||||
| **macOS (Intel)** | [Auto-Claude-2.7.2-beta.10-darwin-x64.dmg](https://github.com/AndyMik90/Auto-Claude/releases/download/v2.7.2-beta.10/Auto-Claude-2.7.2-beta.10-darwin-x64.dmg) |
|
||||
| **Linux** | [Auto-Claude-2.7.2-beta.10-linux-x86_64.AppImage](https://github.com/AndyMik90/Auto-Claude/releases/download/v2.7.2-beta.10/Auto-Claude-2.7.2-beta.10-linux-x86_64.AppImage) |
|
||||
| **Linux (Debian)** | [Auto-Claude-2.7.2-beta.10-linux-amd64.deb](https://github.com/AndyMik90/Auto-Claude/releases/download/v2.7.2-beta.10/Auto-Claude-2.7.2-beta.10-linux-amd64.deb) |
|
||||
| **Linux (Flatpak)** | [Auto-Claude-2.7.2-beta.10-linux-x86_64.flatpak](https://github.com/AndyMik90/Auto-Claude/releases/download/v2.7.2-beta.10/Auto-Claude-2.7.2-beta.10-linux-x86_64.flatpak) |
|
||||
<!-- BETA_DOWNLOADS_END -->
|
||||
|
||||
> All releases include SHA256 checksums and VirusTotal scan results for security verification.
|
||||
|
||||
@@ -19,5 +19,5 @@ Quick Start:
|
||||
See README.md for full documentation.
|
||||
"""
|
||||
|
||||
__version__ = "2.7.6-beta.2"
|
||||
__version__ = "2.7.5"
|
||||
__author__ = "Auto Claude Team"
|
||||
|
||||
@@ -17,14 +17,19 @@ Key Design:
|
||||
|
||||
from __future__ import annotations
|
||||
|
||||
import asyncio
|
||||
import hashlib
|
||||
import logging
|
||||
import os
|
||||
from collections import defaultdict
|
||||
from dataclasses import dataclass
|
||||
from pathlib import Path
|
||||
from typing import Any
|
||||
|
||||
from claude_agent_sdk import AgentDefinition
|
||||
# Note: AgentDefinition import kept for backwards compatibility but no longer used
|
||||
# The Task tool's custom subagent_type feature is broken in Claude Code CLI
|
||||
# See: https://github.com/anthropics/claude-code/issues/8697
|
||||
from claude_agent_sdk import AgentDefinition # noqa: F401
|
||||
|
||||
try:
|
||||
from ...core.client import create_client
|
||||
@@ -48,6 +53,7 @@ try:
|
||||
AgentAgreement,
|
||||
FindingValidationResponse,
|
||||
ParallelOrchestratorResponse,
|
||||
SpecialistResponse,
|
||||
)
|
||||
from .sdk_utils import process_sdk_stream
|
||||
except (ImportError, ValueError, SystemError):
|
||||
@@ -72,10 +78,56 @@ except (ImportError, ValueError, SystemError):
|
||||
AgentAgreement,
|
||||
FindingValidationResponse,
|
||||
ParallelOrchestratorResponse,
|
||||
SpecialistResponse,
|
||||
)
|
||||
from services.sdk_utils import process_sdk_stream
|
||||
|
||||
|
||||
# =============================================================================
|
||||
# Specialist Configuration for Parallel SDK Sessions
|
||||
# =============================================================================
|
||||
|
||||
|
||||
@dataclass
|
||||
class SpecialistConfig:
|
||||
"""Configuration for a specialist agent in parallel SDK sessions."""
|
||||
|
||||
name: str
|
||||
prompt_file: str
|
||||
tools: list[str]
|
||||
description: str
|
||||
|
||||
|
||||
# Define specialist configurations
|
||||
# Each specialist runs as its own SDK session with its own system prompt and tools
|
||||
SPECIALIST_CONFIGS: list[SpecialistConfig] = [
|
||||
SpecialistConfig(
|
||||
name="security",
|
||||
prompt_file="pr_security_agent.md",
|
||||
tools=["Read", "Grep", "Glob"],
|
||||
description="Security vulnerabilities, OWASP Top 10, auth issues, injection, XSS",
|
||||
),
|
||||
SpecialistConfig(
|
||||
name="quality",
|
||||
prompt_file="pr_quality_agent.md",
|
||||
tools=["Read", "Grep", "Glob"],
|
||||
description="Code quality, complexity, duplication, error handling, patterns",
|
||||
),
|
||||
SpecialistConfig(
|
||||
name="logic",
|
||||
prompt_file="pr_logic_agent.md",
|
||||
tools=["Read", "Grep", "Glob"],
|
||||
description="Logic correctness, edge cases, algorithms, race conditions",
|
||||
),
|
||||
SpecialistConfig(
|
||||
name="codebase-fit",
|
||||
prompt_file="pr_codebase_fit_agent.md",
|
||||
tools=["Read", "Grep", "Glob"],
|
||||
description="Naming conventions, ecosystem fit, architectural alignment",
|
||||
),
|
||||
]
|
||||
|
||||
|
||||
logger = logging.getLogger(__name__)
|
||||
|
||||
# Check if debug mode is enabled
|
||||
@@ -335,6 +387,302 @@ class ParallelOrchestratorReviewer:
|
||||
),
|
||||
}
|
||||
|
||||
# =========================================================================
|
||||
# Parallel SDK Sessions Implementation
|
||||
# =========================================================================
|
||||
# This replaces the broken Task tool subagent approach.
|
||||
# Each specialist runs as its own SDK session in parallel via asyncio.gather()
|
||||
# See: https://github.com/anthropics/claude-code/issues/8697
|
||||
|
||||
def _build_specialist_prompt(
|
||||
self,
|
||||
config: SpecialistConfig,
|
||||
context: PRContext,
|
||||
project_root: Path,
|
||||
) -> str:
|
||||
"""Build the full prompt for a specialist agent.
|
||||
|
||||
Args:
|
||||
config: Specialist configuration
|
||||
context: PR context with files and patches
|
||||
project_root: Working directory for the agent
|
||||
|
||||
Returns:
|
||||
Full system prompt with context injected
|
||||
"""
|
||||
# Load base prompt from file
|
||||
base_prompt = self._load_prompt(config.prompt_file)
|
||||
if not base_prompt:
|
||||
base_prompt = f"You are a {config.name} specialist for PR review."
|
||||
|
||||
# Inject working directory using the existing helper
|
||||
with_working_dir = create_working_dir_injector(project_root)
|
||||
prompt_with_cwd = with_working_dir(
|
||||
base_prompt,
|
||||
f"You are a {config.name} specialist. Find {config.description}.",
|
||||
)
|
||||
|
||||
# Build file list
|
||||
files_list = []
|
||||
for file in context.changed_files:
|
||||
files_list.append(
|
||||
f"- `{file.path}` (+{file.additions}/-{file.deletions}) - {file.status}"
|
||||
)
|
||||
|
||||
# Build diff content (limited to avoid context overflow)
|
||||
patches = []
|
||||
MAX_DIFF_CHARS = 150_000 # Smaller limit per specialist
|
||||
|
||||
for file in context.changed_files:
|
||||
if file.patch:
|
||||
patches.append(f"\n### File: {file.path}\n{file.patch}")
|
||||
|
||||
diff_content = "\n".join(patches)
|
||||
if len(diff_content) > MAX_DIFF_CHARS:
|
||||
diff_content = diff_content[:MAX_DIFF_CHARS] + "\n\n... (diff truncated)"
|
||||
|
||||
# Compose full prompt with PR context
|
||||
pr_context = f"""
|
||||
## PR Context
|
||||
|
||||
**PR #{context.pr_number}**: {context.title}
|
||||
|
||||
**Description:**
|
||||
{context.description or "(No description provided)"}
|
||||
|
||||
### Changed Files ({len(context.changed_files)} files, +{context.total_additions}/-{context.total_deletions})
|
||||
{chr(10).join(files_list)}
|
||||
|
||||
### Diff
|
||||
{diff_content}
|
||||
|
||||
## Your Task
|
||||
|
||||
Analyze this PR for {config.description}.
|
||||
Use the Read, Grep, and Glob tools to explore the codebase as needed.
|
||||
Report findings with specific file paths, line numbers, and code evidence.
|
||||
"""
|
||||
|
||||
return prompt_with_cwd + pr_context
|
||||
|
||||
async def _run_specialist_session(
|
||||
self,
|
||||
config: SpecialistConfig,
|
||||
context: PRContext,
|
||||
project_root: Path,
|
||||
model: str,
|
||||
thinking_budget: int | None,
|
||||
) -> tuple[str, list[PRReviewFinding]]:
|
||||
"""Run a single specialist as its own SDK session.
|
||||
|
||||
Args:
|
||||
config: Specialist configuration
|
||||
context: PR context
|
||||
project_root: Working directory
|
||||
model: Model to use
|
||||
thinking_budget: Max thinking tokens
|
||||
|
||||
Returns:
|
||||
Tuple of (specialist_name, findings)
|
||||
"""
|
||||
safe_print(
|
||||
f"[Specialist:{config.name}] Starting analysis...",
|
||||
flush=True,
|
||||
)
|
||||
|
||||
# Build the specialist prompt with PR context
|
||||
prompt = self._build_specialist_prompt(config, context, project_root)
|
||||
|
||||
try:
|
||||
# Create SDK client for this specialist
|
||||
# Note: Agent type uses the generic "pr_reviewer" since individual
|
||||
# specialist types aren't registered in AGENT_CONFIGS. The specialist-specific
|
||||
# system prompt handles differentiation.
|
||||
client = create_client(
|
||||
project_dir=project_root,
|
||||
spec_dir=self.github_dir,
|
||||
model=model,
|
||||
agent_type="pr_reviewer",
|
||||
max_thinking_tokens=thinking_budget,
|
||||
output_format={
|
||||
"type": "json_schema",
|
||||
"schema": SpecialistResponse.model_json_schema(),
|
||||
},
|
||||
)
|
||||
|
||||
async with client:
|
||||
await client.query(prompt)
|
||||
|
||||
# Process SDK stream
|
||||
stream_result = await process_sdk_stream(
|
||||
client=client,
|
||||
context_name=f"Specialist:{config.name}",
|
||||
model=model,
|
||||
system_prompt=prompt,
|
||||
agent_definitions={}, # No subagents for specialists
|
||||
)
|
||||
|
||||
error = stream_result.get("error")
|
||||
if error:
|
||||
logger.error(f"[Specialist:{config.name}] SDK stream failed: {error}")
|
||||
safe_print(
|
||||
f"[Specialist:{config.name}] Analysis failed: {error}",
|
||||
flush=True,
|
||||
)
|
||||
return (config.name, [])
|
||||
|
||||
# Parse structured output
|
||||
structured_output = stream_result.get("structured_output")
|
||||
findings = self._parse_specialist_output(
|
||||
config.name, structured_output, stream_result.get("result_text", "")
|
||||
)
|
||||
|
||||
safe_print(
|
||||
f"[Specialist:{config.name}] Complete: {len(findings)} findings",
|
||||
flush=True,
|
||||
)
|
||||
|
||||
return (config.name, findings)
|
||||
|
||||
except Exception as e:
|
||||
logger.error(
|
||||
f"[Specialist:{config.name}] Session failed: {e}",
|
||||
exc_info=True,
|
||||
)
|
||||
safe_print(
|
||||
f"[Specialist:{config.name}] Error: {e}",
|
||||
flush=True,
|
||||
)
|
||||
return (config.name, [])
|
||||
|
||||
def _parse_specialist_output(
|
||||
self,
|
||||
specialist_name: str,
|
||||
structured_output: dict[str, Any] | None,
|
||||
result_text: str,
|
||||
) -> list[PRReviewFinding]:
|
||||
"""Parse findings from specialist output.
|
||||
|
||||
Args:
|
||||
specialist_name: Name of the specialist
|
||||
structured_output: Structured JSON output if available
|
||||
result_text: Raw text output as fallback
|
||||
|
||||
Returns:
|
||||
List of PRReviewFinding objects
|
||||
"""
|
||||
findings = []
|
||||
|
||||
if structured_output:
|
||||
try:
|
||||
result = SpecialistResponse.model_validate(structured_output)
|
||||
|
||||
for f in result.findings:
|
||||
finding_id = hashlib.md5(
|
||||
f"{f.file}:{f.line}:{f.title}".encode(),
|
||||
usedforsecurity=False,
|
||||
).hexdigest()[:12]
|
||||
|
||||
category = map_category(f.category)
|
||||
|
||||
try:
|
||||
severity = ReviewSeverity(f.severity.lower())
|
||||
except ValueError:
|
||||
severity = ReviewSeverity.MEDIUM
|
||||
|
||||
finding = PRReviewFinding(
|
||||
id=finding_id,
|
||||
file=f.file,
|
||||
line=f.line,
|
||||
end_line=f.end_line,
|
||||
title=f.title,
|
||||
description=f.description,
|
||||
category=category,
|
||||
severity=severity,
|
||||
suggested_fix=f.suggested_fix or "",
|
||||
evidence=f.evidence,
|
||||
source_agents=[specialist_name],
|
||||
is_impact_finding=f.is_impact_finding,
|
||||
)
|
||||
findings.append(finding)
|
||||
|
||||
logger.info(
|
||||
f"[Specialist:{specialist_name}] Parsed {len(findings)} findings from structured output"
|
||||
)
|
||||
|
||||
except Exception as e:
|
||||
logger.error(
|
||||
f"[Specialist:{specialist_name}] Failed to parse structured output: {e}"
|
||||
)
|
||||
# Fall through to text parsing
|
||||
|
||||
if not findings and result_text:
|
||||
# Fallback to text parsing
|
||||
findings = self._parse_text_output(result_text)
|
||||
for f in findings:
|
||||
f.source_agents = [specialist_name]
|
||||
|
||||
return findings
|
||||
|
||||
async def _run_parallel_specialists(
|
||||
self,
|
||||
context: PRContext,
|
||||
project_root: Path,
|
||||
model: str,
|
||||
thinking_budget: int | None,
|
||||
) -> tuple[list[PRReviewFinding], list[str]]:
|
||||
"""Run all specialists in parallel and collect findings.
|
||||
|
||||
Args:
|
||||
context: PR context
|
||||
project_root: Working directory
|
||||
model: Model to use
|
||||
thinking_budget: Max thinking tokens
|
||||
|
||||
Returns:
|
||||
Tuple of (all_findings, agents_invoked)
|
||||
"""
|
||||
safe_print(
|
||||
f"[ParallelOrchestrator] Launching {len(SPECIALIST_CONFIGS)} specialists in parallel...",
|
||||
flush=True,
|
||||
)
|
||||
|
||||
# Create tasks for all specialists
|
||||
tasks = [
|
||||
self._run_specialist_session(
|
||||
config=config,
|
||||
context=context,
|
||||
project_root=project_root,
|
||||
model=model,
|
||||
thinking_budget=thinking_budget,
|
||||
)
|
||||
for config in SPECIALIST_CONFIGS
|
||||
]
|
||||
|
||||
# Run all specialists in parallel
|
||||
results = await asyncio.gather(*tasks, return_exceptions=True)
|
||||
|
||||
# Collect findings and track which agents ran
|
||||
all_findings: list[PRReviewFinding] = []
|
||||
agents_invoked: list[str] = []
|
||||
|
||||
for result in results:
|
||||
if isinstance(result, Exception):
|
||||
logger.error(f"[ParallelOrchestrator] Specialist task failed: {result}")
|
||||
continue
|
||||
|
||||
specialist_name, findings = result
|
||||
agents_invoked.append(specialist_name)
|
||||
all_findings.extend(findings)
|
||||
|
||||
safe_print(
|
||||
f"[ParallelOrchestrator] All specialists complete. "
|
||||
f"Total findings: {len(all_findings)}",
|
||||
flush=True,
|
||||
)
|
||||
|
||||
return (all_findings, agents_invoked)
|
||||
|
||||
def _build_orchestrator_prompt(self, context: PRContext) -> str:
|
||||
"""Build full prompt for orchestrator with PR context."""
|
||||
# Load orchestrator prompt
|
||||
@@ -720,11 +1068,6 @@ The SDK will run invoked agents in parallel automatically.
|
||||
# LLM agents now discover relevant files themselves via Read, Grep, Glob tools
|
||||
# No need to pre-scan the codebase programmatically
|
||||
|
||||
# Build orchestrator prompt AFTER worktree creation and related files rescan
|
||||
prompt = self._build_orchestrator_prompt(context)
|
||||
# Capture agent definitions for debug logging (with worktree path)
|
||||
agent_defs = self._define_specialist_agents(project_root)
|
||||
|
||||
# Use model and thinking level from config (user settings)
|
||||
# Resolve model shorthand via environment variable override if configured
|
||||
model_shorthand = self.config.model or "sonnet"
|
||||
@@ -737,122 +1080,114 @@ The SDK will run invoked agents in parallel automatically.
|
||||
f"thinking_level={thinking_level}, thinking_budget={thinking_budget}"
|
||||
)
|
||||
|
||||
# Create client with subagents defined
|
||||
# SDK handles parallel execution when Claude invokes multiple Task tools
|
||||
client = self._create_sdk_client(project_root, model, thinking_budget)
|
||||
|
||||
self._report_progress(
|
||||
"orchestrating",
|
||||
40,
|
||||
"Orchestrator delegating to specialist agents...",
|
||||
"Running specialist agents in parallel...",
|
||||
pr_number=context.pr_number,
|
||||
)
|
||||
|
||||
# Run orchestrator session using shared SDK stream processor
|
||||
# Retry logic for tool use concurrency errors
|
||||
MAX_RETRIES = 3
|
||||
RETRY_DELAY = 2.0 # seconds between retries
|
||||
# =================================================================
|
||||
# PARALLEL SDK SESSIONS APPROACH
|
||||
# =================================================================
|
||||
# Instead of using broken Task tool subagents, we spawn each
|
||||
# specialist as its own SDK session and run them in parallel.
|
||||
# See: https://github.com/anthropics/claude-code/issues/8697
|
||||
#
|
||||
# This gives us:
|
||||
# - True parallel execution via asyncio.gather()
|
||||
# - Full control over each specialist's tools and prompts
|
||||
# - No dependency on broken CLI features
|
||||
# =================================================================
|
||||
|
||||
result_text = ""
|
||||
structured_output = None
|
||||
agents_invoked = []
|
||||
msg_count = 0
|
||||
last_error = None
|
||||
# Run all specialists in parallel
|
||||
findings, agents_invoked = await self._run_parallel_specialists(
|
||||
context=context,
|
||||
project_root=project_root,
|
||||
model=model,
|
||||
thinking_budget=thinking_budget,
|
||||
)
|
||||
|
||||
for attempt in range(MAX_RETRIES):
|
||||
if attempt > 0:
|
||||
logger.info(
|
||||
f"[ParallelOrchestrator] Retry attempt {attempt}/{MAX_RETRIES - 1} "
|
||||
f"after tool concurrency error"
|
||||
)
|
||||
safe_print(
|
||||
f"[ParallelOrchestrator] Retry {attempt}/{MAX_RETRIES - 1} "
|
||||
f"(tool concurrency error detected)"
|
||||
)
|
||||
# Small delay before retry
|
||||
import asyncio
|
||||
# Log results
|
||||
logger.info(
|
||||
f"[ParallelOrchestrator] Parallel specialists complete: "
|
||||
f"{len(findings)} findings from {len(agents_invoked)} agents"
|
||||
)
|
||||
|
||||
await asyncio.sleep(RETRY_DELAY)
|
||||
# Skip the old orchestrator session code - findings come from parallel specialists
|
||||
# The code below (structured output parsing, retries, etc.) is no longer needed
|
||||
# as _run_parallel_specialists handles everything
|
||||
|
||||
# Recreate client for retry (fresh session)
|
||||
client = self._create_sdk_client(
|
||||
project_root, model, thinking_budget
|
||||
)
|
||||
# NOTE: The following block is kept but skipped via this marker
|
||||
if False: # DISABLED: Old orchestrator + Task tool approach
|
||||
# Old code for reference - to be removed after testing
|
||||
prompt = self._build_orchestrator_prompt(context)
|
||||
agent_defs = self._define_specialist_agents(project_root)
|
||||
client = self._create_sdk_client(project_root, model, thinking_budget)
|
||||
|
||||
try:
|
||||
async with client:
|
||||
await client.query(prompt)
|
||||
MAX_RETRIES = 3
|
||||
RETRY_DELAY = 2.0
|
||||
|
||||
result_text = ""
|
||||
structured_output = None
|
||||
msg_count = 0
|
||||
last_error = None
|
||||
|
||||
for attempt in range(MAX_RETRIES):
|
||||
if attempt > 0:
|
||||
logger.info(
|
||||
f"[ParallelOrchestrator] Retry attempt {attempt}/{MAX_RETRIES - 1} "
|
||||
f"after tool concurrency error"
|
||||
)
|
||||
safe_print(
|
||||
f"[ParallelOrchestrator] Running orchestrator ({model})...",
|
||||
flush=True,
|
||||
f"[ParallelOrchestrator] Retry {attempt}/{MAX_RETRIES - 1} "
|
||||
f"(tool concurrency error detected)"
|
||||
)
|
||||
await asyncio.sleep(RETRY_DELAY)
|
||||
client = self._create_sdk_client(
|
||||
project_root, model, thinking_budget
|
||||
)
|
||||
|
||||
# Process SDK stream with shared utility
|
||||
stream_result = await process_sdk_stream(
|
||||
client=client,
|
||||
context_name="ParallelOrchestrator",
|
||||
model=model,
|
||||
system_prompt=prompt,
|
||||
agent_definitions=agent_defs,
|
||||
)
|
||||
try:
|
||||
async with client:
|
||||
await client.query(prompt)
|
||||
|
||||
error = stream_result.get("error")
|
||||
|
||||
# Check for tool concurrency error specifically
|
||||
if (
|
||||
error == "tool_use_concurrency_error"
|
||||
and attempt < MAX_RETRIES - 1
|
||||
):
|
||||
logger.warning(
|
||||
f"[ParallelOrchestrator] Tool concurrency error on attempt {attempt + 1}, "
|
||||
f"will retry..."
|
||||
safe_print(
|
||||
f"[ParallelOrchestrator] Running orchestrator ({model})...",
|
||||
flush=True,
|
||||
)
|
||||
last_error = error
|
||||
continue # Retry
|
||||
|
||||
# Check for other stream processing errors
|
||||
if error:
|
||||
logger.error(
|
||||
f"[ParallelOrchestrator] SDK stream failed: {error}"
|
||||
stream_result = await process_sdk_stream(
|
||||
client=client,
|
||||
context_name="ParallelOrchestrator",
|
||||
model=model,
|
||||
system_prompt=prompt,
|
||||
agent_definitions=agent_defs,
|
||||
)
|
||||
raise RuntimeError(f"SDK stream processing failed: {error}")
|
||||
|
||||
# Success - extract results and break retry loop
|
||||
result_text = stream_result["result_text"]
|
||||
structured_output = stream_result["structured_output"]
|
||||
agents_invoked = stream_result["agents_invoked"]
|
||||
msg_count = stream_result["msg_count"]
|
||||
break # Success, exit retry loop
|
||||
error = stream_result.get("error")
|
||||
|
||||
except Exception as e:
|
||||
# Check if this is a retryable error
|
||||
error_str = str(e).lower()
|
||||
is_retryable = (
|
||||
"400" in error_str
|
||||
or "concurrency" in error_str
|
||||
or "tool_use" in error_str
|
||||
)
|
||||
if (
|
||||
error == "tool_use_concurrency_error"
|
||||
and attempt < MAX_RETRIES - 1
|
||||
):
|
||||
last_error = error
|
||||
continue
|
||||
if error:
|
||||
raise RuntimeError(f"SDK stream processing failed: {error}")
|
||||
result_text = stream_result["result_text"]
|
||||
structured_output = stream_result["structured_output"]
|
||||
agents_invoked = stream_result["agents_invoked"]
|
||||
break
|
||||
except Exception as e:
|
||||
if attempt < MAX_RETRIES - 1:
|
||||
last_error = str(e)
|
||||
continue
|
||||
raise
|
||||
else:
|
||||
raise RuntimeError(f"Orchestrator failed after {MAX_RETRIES} attempts")
|
||||
|
||||
if is_retryable and attempt < MAX_RETRIES - 1:
|
||||
logger.warning(
|
||||
f"[ParallelOrchestrator] Retryable error on attempt {attempt + 1}: {e}"
|
||||
)
|
||||
last_error = str(e)
|
||||
continue # Retry
|
||||
|
||||
# Not retryable or out of retries - re-raise
|
||||
raise
|
||||
else:
|
||||
# All retries exhausted
|
||||
logger.error(
|
||||
f"[ParallelOrchestrator] Failed after {MAX_RETRIES} attempts. "
|
||||
f"Last error: {last_error}"
|
||||
)
|
||||
raise RuntimeError(
|
||||
f"Orchestrator failed after {MAX_RETRIES} retry attempts. "
|
||||
f"Last error: {last_error}"
|
||||
)
|
||||
# END DISABLED BLOCK
|
||||
|
||||
self._report_progress(
|
||||
"finalizing",
|
||||
@@ -861,20 +1196,9 @@ The SDK will run invoked agents in parallel automatically.
|
||||
pr_number=context.pr_number,
|
||||
)
|
||||
|
||||
# Parse findings from output (structured output also returns agents)
|
||||
findings, agents_from_structured = self._extract_structured_output(
|
||||
structured_output, result_text
|
||||
)
|
||||
|
||||
# Use agents from structured output (more reliable than streaming detection)
|
||||
final_agents = (
|
||||
agents_from_structured if agents_from_structured else agents_invoked
|
||||
)
|
||||
logger.info(
|
||||
f"[ParallelOrchestrator] Session complete. Agents invoked: {final_agents}"
|
||||
)
|
||||
# Log completion with agent info
|
||||
safe_print(
|
||||
f"[ParallelOrchestrator] Complete. Agents invoked: {final_agents}",
|
||||
f"[ParallelOrchestrator] Complete. Agents invoked: {agents_invoked}",
|
||||
flush=True,
|
||||
)
|
||||
|
||||
@@ -977,7 +1301,7 @@ The SDK will run invoked agents in parallel automatically.
|
||||
verdict_reasoning=verdict_reasoning,
|
||||
blockers=blockers,
|
||||
findings=unique_findings,
|
||||
agents_invoked=final_agents,
|
||||
agents_invoked=agents_invoked,
|
||||
)
|
||||
|
||||
# Map verdict to overall_status
|
||||
|
||||
@@ -537,6 +537,53 @@ class ValidationSummary(BaseModel):
|
||||
)
|
||||
|
||||
|
||||
class SpecialistFinding(BaseModel):
|
||||
"""A finding from a specialist agent (used in parallel SDK sessions)."""
|
||||
|
||||
severity: Literal["critical", "high", "medium", "low"] = Field(
|
||||
description="Issue severity level"
|
||||
)
|
||||
category: Literal[
|
||||
"security", "quality", "logic", "performance", "pattern", "test", "docs"
|
||||
] = Field(description="Issue category")
|
||||
title: str = Field(description="Brief issue title (max 80 chars)")
|
||||
description: str = Field(description="Detailed explanation of the issue")
|
||||
file: str = Field(description="File path where issue was found")
|
||||
line: int = Field(0, description="Line number of the issue")
|
||||
end_line: int | None = Field(None, description="End line number if multi-line")
|
||||
suggested_fix: str | None = Field(None, description="How to fix this issue")
|
||||
evidence: str = Field(
|
||||
min_length=1,
|
||||
description="Actual code snippet examined that shows the issue. Required.",
|
||||
)
|
||||
is_impact_finding: bool = Field(
|
||||
False,
|
||||
description="True if this is about affected code outside the PR (callers, dependencies)",
|
||||
)
|
||||
|
||||
|
||||
class SpecialistResponse(BaseModel):
|
||||
"""Response schema for individual specialist agent (parallel SDK sessions).
|
||||
|
||||
Used when each specialist runs as its own SDK session rather than via Task tool.
|
||||
"""
|
||||
|
||||
specialist_name: str = Field(
|
||||
description="Name of the specialist (security, quality, logic, codebase-fit)"
|
||||
)
|
||||
analysis_summary: str = Field(
|
||||
description="Brief summary of what was analyzed"
|
||||
)
|
||||
files_examined: list[str] = Field(
|
||||
default_factory=list,
|
||||
description="List of files that were examined",
|
||||
)
|
||||
findings: list[SpecialistFinding] = Field(
|
||||
default_factory=list,
|
||||
description="Issues found during analysis",
|
||||
)
|
||||
|
||||
|
||||
class ParallelOrchestratorResponse(BaseModel):
|
||||
"""Complete response schema for parallel orchestrator PR review."""
|
||||
|
||||
|
||||
+10
-35
@@ -1,6 +1,6 @@
|
||||
{
|
||||
"name": "auto-claude-ui",
|
||||
"version": "2.7.6-beta.2",
|
||||
"version": "2.7.5",
|
||||
"type": "module",
|
||||
"description": "Desktop UI for Auto Claude autonomous coding framework",
|
||||
"homepage": "https://github.com/AndyMik90/Auto-Claude",
|
||||
@@ -36,8 +36,6 @@
|
||||
"package:win": "node scripts/package-with-python.cjs --win",
|
||||
"package:linux": "node scripts/package-with-python.cjs --linux",
|
||||
"package:flatpak": "node scripts/package-with-python.cjs --linux flatpak",
|
||||
"verify:linux": "node scripts/verify-linux-packages.cjs dist",
|
||||
"test:verify-linux": "node --test scripts/verify-linux-packages.test.mjs",
|
||||
"start:packaged:mac": "open dist/mac-arm64/Auto-Claude.app || open dist/mac/Auto-Claude.app",
|
||||
"start:packaged:win": "start \"\" \"dist\\win-unpacked\\Auto-Claude.exe\"",
|
||||
"start:packaged:linux": "./dist/linux-unpacked/auto-claude",
|
||||
@@ -100,7 +98,6 @@
|
||||
"semver": "^7.7.3",
|
||||
"tailwind-merge": "^3.4.0",
|
||||
"uuid": "^13.0.0",
|
||||
"xstate": "^5.26.0",
|
||||
"zod": "^4.2.1",
|
||||
"zustand": "^5.0.9"
|
||||
},
|
||||
@@ -168,6 +165,14 @@
|
||||
"from": "resources/icon.ico",
|
||||
"to": "icon.ico"
|
||||
},
|
||||
{
|
||||
"from": "python-runtime/${os}-${arch}/python",
|
||||
"to": "python"
|
||||
},
|
||||
{
|
||||
"from": "python-runtime/${os}-${arch}/site-packages",
|
||||
"to": "python-site-packages"
|
||||
},
|
||||
{
|
||||
"from": "../backend",
|
||||
"to": "backend",
|
||||
@@ -197,16 +202,6 @@
|
||||
"target": [
|
||||
"dmg",
|
||||
"zip"
|
||||
],
|
||||
"extraResources": [
|
||||
{
|
||||
"from": "python-runtime/${os}-${arch}/python",
|
||||
"to": "python"
|
||||
},
|
||||
{
|
||||
"from": "python-runtime/${os}-${arch}/site-packages",
|
||||
"to": "python-site-packages"
|
||||
}
|
||||
]
|
||||
},
|
||||
"win": {
|
||||
@@ -214,16 +209,6 @@
|
||||
"target": [
|
||||
"nsis",
|
||||
"zip"
|
||||
],
|
||||
"extraResources": [
|
||||
{
|
||||
"from": "python-runtime/${os}-${arch}/python",
|
||||
"to": "python"
|
||||
},
|
||||
{
|
||||
"from": "python-runtime/${os}-${arch}/site-packages",
|
||||
"to": "python-site-packages"
|
||||
}
|
||||
]
|
||||
},
|
||||
"linux": {
|
||||
@@ -233,17 +218,7 @@
|
||||
"deb",
|
||||
"flatpak"
|
||||
],
|
||||
"category": "Development",
|
||||
"extraResources": [
|
||||
{
|
||||
"from": "python-runtime/${os}-${arch}/python",
|
||||
"to": "python"
|
||||
},
|
||||
{
|
||||
"from": "python-runtime/${os}-${arch}/site-packages",
|
||||
"to": "python-site-packages"
|
||||
}
|
||||
]
|
||||
"category": "Development"
|
||||
},
|
||||
"flatpak": {
|
||||
"runtime": "org.freedesktop.Platform",
|
||||
|
||||
@@ -870,6 +870,16 @@ function parseLogLine(line: string): { source: string; content: string; isError:
|
||||
};
|
||||
}
|
||||
|
||||
// Check for parallel SDK specialist logs (Specialist:name format)
|
||||
const specialistMatch = line.match(/^\[Specialist:([\w-]+)\]\s*(.*)$/);
|
||||
if (specialistMatch) {
|
||||
return {
|
||||
source: `Specialist:${specialistMatch[1]}`,
|
||||
content: specialistMatch[2],
|
||||
isError: false,
|
||||
};
|
||||
}
|
||||
|
||||
for (const pattern of patterns) {
|
||||
const match = line.match(pattern);
|
||||
if (match) {
|
||||
@@ -970,8 +980,9 @@ function getPhaseFromSource(source: string): PRLogPhase {
|
||||
|
||||
if (contextSources.includes(source)) return "context";
|
||||
if (analysisSources.includes(source)) return "analysis";
|
||||
// Specialist agents (Agent:xxx) are part of analysis phase
|
||||
// Specialist agents (Agent:xxx and Specialist:xxx) are part of analysis phase
|
||||
if (source.startsWith("Agent:")) return "analysis";
|
||||
if (source.startsWith("Specialist:")) return "analysis";
|
||||
if (synthesisSources.includes(source)) return "synthesis";
|
||||
return "synthesis"; // Default to synthesis for unknown sources
|
||||
}
|
||||
|
||||
@@ -61,7 +61,7 @@ const SOURCE_COLORS: Record<string, string> = {
|
||||
'Progress': 'bg-green-500/20 text-green-400',
|
||||
'PR Review Engine': 'bg-indigo-500/20 text-indigo-400',
|
||||
'Summary': 'bg-emerald-500/20 text-emerald-400',
|
||||
// Specialist agents (from parallel orchestrator)
|
||||
// Specialist agents (from parallel orchestrator - old Task tool approach)
|
||||
'Agent:logic-reviewer': 'bg-blue-600/20 text-blue-400',
|
||||
'Agent:quality-reviewer': 'bg-indigo-600/20 text-indigo-400',
|
||||
'Agent:security-reviewer': 'bg-red-600/20 text-red-400',
|
||||
@@ -70,6 +70,11 @@ const SOURCE_COLORS: Record<string, string> = {
|
||||
'Agent:resolution-verifier': 'bg-teal-600/20 text-teal-400',
|
||||
'Agent:new-code-reviewer': 'bg-cyan-600/20 text-cyan-400',
|
||||
'Agent:comment-analyzer': 'bg-gray-500/20 text-gray-400',
|
||||
// Parallel SDK specialists (new approach using parallel SDK sessions)
|
||||
'Specialist:security': 'bg-red-600/20 text-red-400',
|
||||
'Specialist:quality': 'bg-indigo-600/20 text-indigo-400',
|
||||
'Specialist:logic': 'bg-blue-600/20 text-blue-400',
|
||||
'Specialist:codebase-fit': 'bg-emerald-600/20 text-emerald-400',
|
||||
'default': 'bg-muted text-muted-foreground'
|
||||
};
|
||||
|
||||
@@ -107,8 +112,8 @@ function groupEntriesByAgent(entries: PRLogEntry[]): {
|
||||
const otherEntries: PRLogEntry[] = [];
|
||||
|
||||
for (const entry of entries) {
|
||||
if (entry.source?.startsWith('Agent:')) {
|
||||
// Agent results
|
||||
if (entry.source?.startsWith('Agent:') || entry.source?.startsWith('Specialist:')) {
|
||||
// Agent/Specialist results (both old Task tool and new parallel SDK approaches)
|
||||
const existing = agentMap.get(entry.source) || [];
|
||||
existing.push(entry);
|
||||
agentMap.set(entry.source, existing);
|
||||
@@ -451,15 +456,66 @@ interface AgentLogGroupProps {
|
||||
onToggle: () => void;
|
||||
}
|
||||
|
||||
// Patterns that are uninteresting as summary entries
|
||||
const SKIP_AS_SUMMARY_PATTERNS = [
|
||||
/^Starting analysis\.\.\.$/,
|
||||
/^Processing SDK stream\.\.\.$/,
|
||||
/^Processing\.\.\./,
|
||||
/^Awaiting response stream\.\.\.$/,
|
||||
];
|
||||
|
||||
function isBoringSummary(content: string): boolean {
|
||||
return SKIP_AS_SUMMARY_PATTERNS.some(pattern => pattern.test(content));
|
||||
}
|
||||
|
||||
// Find a meaningful summary entry - skip boring entries and prefer "AI response" or "Complete"
|
||||
function findSummaryEntry(entries: PRLogEntry[]): { summaryEntry: PRLogEntry | undefined; otherEntries: PRLogEntry[] } {
|
||||
if (entries.length === 0) return { summaryEntry: undefined, otherEntries: [] };
|
||||
|
||||
// Look for the most informative entry to show as summary
|
||||
// Priority: 1) "Complete:" entry, 2) "AI response:" entry, 3) first non-boring entry
|
||||
const completeEntry = entries.find(e => e.content.startsWith('Complete:'));
|
||||
if (completeEntry) {
|
||||
return {
|
||||
summaryEntry: completeEntry,
|
||||
otherEntries: entries.filter(e => e !== completeEntry),
|
||||
};
|
||||
}
|
||||
|
||||
const aiResponseEntry = entries.find(e => e.content.startsWith('AI response:'));
|
||||
if (aiResponseEntry) {
|
||||
return {
|
||||
summaryEntry: aiResponseEntry,
|
||||
otherEntries: entries.filter(e => e !== aiResponseEntry),
|
||||
};
|
||||
}
|
||||
|
||||
// Find first non-boring entry
|
||||
const meaningfulEntry = entries.find(e => !isBoringSummary(e.content));
|
||||
if (meaningfulEntry) {
|
||||
return {
|
||||
summaryEntry: meaningfulEntry,
|
||||
otherEntries: entries.filter(e => e !== meaningfulEntry),
|
||||
};
|
||||
}
|
||||
|
||||
// Fallback to first entry
|
||||
return {
|
||||
summaryEntry: entries[0],
|
||||
otherEntries: entries.slice(1),
|
||||
};
|
||||
}
|
||||
|
||||
function AgentLogGroup({ group, isExpanded, onToggle }: AgentLogGroupProps) {
|
||||
const { t } = useTranslation(['common']);
|
||||
const { agentName, entries } = group;
|
||||
const hasMultipleEntries = entries.length > 1;
|
||||
const firstEntry = entries[0];
|
||||
const remainingEntries = entries.slice(1);
|
||||
|
||||
// Extract display name from "Agent:logic-reviewer" -> "logic-reviewer"
|
||||
const displayName = agentName.replace('Agent:', '');
|
||||
// Find a meaningful summary entry instead of just using the first one
|
||||
const { summaryEntry, otherEntries } = findSummaryEntry(entries);
|
||||
const hasMoreEntries = otherEntries.length > 0;
|
||||
|
||||
// Extract display name from "Agent:logic-reviewer" -> "logic-reviewer" or "Specialist:security" -> "security"
|
||||
const displayName = agentName.replace('Agent:', '').replace('Specialist:', '');
|
||||
|
||||
const getSourceColor = (source: string) => {
|
||||
return SOURCE_COLORS[source] || SOURCE_COLORS.default;
|
||||
@@ -477,7 +533,7 @@ function AgentLogGroup({ group, isExpanded, onToggle }: AgentLogGroupProps) {
|
||||
>
|
||||
{displayName}
|
||||
</Badge>
|
||||
{hasMultipleEntries && (
|
||||
{hasMoreEntries && (
|
||||
<button
|
||||
onClick={onToggle}
|
||||
className={cn(
|
||||
@@ -489,28 +545,28 @@ function AgentLogGroup({ group, isExpanded, onToggle }: AgentLogGroupProps) {
|
||||
{isExpanded ? (
|
||||
<>
|
||||
<ChevronDown className="h-3 w-3" />
|
||||
<span>{t('common:prReview.logs.hideMore', { count: remainingEntries.length })}</span>
|
||||
<span>{t('common:prReview.logs.hideMore', { count: otherEntries.length })}</span>
|
||||
</>
|
||||
) : (
|
||||
<>
|
||||
<ChevronRight className="h-3 w-3" />
|
||||
<span>{t('common:prReview.logs.showMore', { count: remainingEntries.length })}</span>
|
||||
<span>{t('common:prReview.logs.showMore', { count: otherEntries.length })}</span>
|
||||
</>
|
||||
)}
|
||||
</button>
|
||||
)}
|
||||
</div>
|
||||
|
||||
{/* First entry (summary) - always visible */}
|
||||
{firstEntry && (
|
||||
<LogEntry entry={{ ...firstEntry, source: undefined }} />
|
||||
{/* Summary entry - always visible (most informative entry, not necessarily first) */}
|
||||
{summaryEntry && (
|
||||
<LogEntry entry={{ ...summaryEntry, source: undefined }} />
|
||||
)}
|
||||
</div>
|
||||
|
||||
{/* Collapsible section for remaining entries */}
|
||||
{hasMultipleEntries && isExpanded && (
|
||||
{/* Collapsible section for other entries */}
|
||||
{hasMoreEntries && isExpanded && (
|
||||
<div className="border-t border-border/30 bg-secondary/10 p-2 space-y-1">
|
||||
{remainingEntries.map((entry, idx) => (
|
||||
{otherEntries.map((entry, idx) => (
|
||||
<LogEntry key={`${entry.timestamp}-${idx}`} entry={{ ...entry, source: undefined }} />
|
||||
))}
|
||||
</div>
|
||||
|
||||
Reference in New Issue
Block a user