mirror of
https://github.com/leonvanzyl/autocoder.git
synced 2026-02-01 23:13:36 +00:00
Token reduction (~40% per session, ~2.3M fewer tokens per 200-feature project): - Agent-type-specific tool lists: coding 9, testing 5, init 5 (was 19 for all) - Right-sized max_turns: coding 300, testing 100 (was 1000 for all) - Trimmed coding prompt template (~150 lines removed) - Streamlined testing prompt with batch support - YOLO mode now strips browser testing instructions from prompt - Added Grep, WebFetch, WebSearch to expand project session Performance improvements: - Rate limit retries start at ~15s with jitter (was fixed 60s) - Post-spawn delay reduced to 0.5s (was 2s) - Orchestrator consolidated to 1 DB query per loop (was 5-7) - Testing agents batch 3 features per session (was 1) - Smart context compaction preserves critical state, discards noise Bug fixes: - Removed ghost feature_release_testing MCP tool (wasted tokens every test session) - Forward all 9 Vertex AI env vars to chat sessions (was missing 3) - Fix DetachedInstanceError risk in test batch ORM access - Prevent duplicate testing of same features in parallel mode Code deduplication: - _get_project_path(): 9 copies -> 1 shared utility (project_helpers.py) - validate_project_name(): 9 copies -> 2 variants in 1 file (validation.py) - ROOT_DIR: 10 copies -> 1 definition (chat_constants.py) - API_ENV_VARS: 4 copies -> 1 source of truth (env_constants.py) Security hardening: - Unified sensitive directory blocklist (14 dirs, was two divergent lists) - Cached get_blocked_paths() for O(1) directory listing checks - Terminal security warning when ALLOW_REMOTE=1 exposes WebSocket - 20 new security tests for EXTRA_READ_PATHS blocking - Extracted _validate_command_list() and _validate_pkill_processes() helpers Type safety: - 87 mypy errors -> 0 across 58 source files - Installed types-PyYAML for proper yaml stub types - Fixed SQLAlchemy Column[T] coercions across all routers Dead code removed: - 13 files deleted (~2,679 lines): unused UI components, debug logs, outdated docs - 7 unused npm packages removed (Radix UI components with 0 imports) - AgentAvatar.tsx reduced from 615 -> 119 lines (SVGs extracted to mascotData.tsx) New CLI options: - --testing-batch-size (1-5) for parallel mode test batching - --testing-feature-ids for direct multi-feature testing Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
269 lines
9.1 KiB
Python
269 lines
9.1 KiB
Python
#!/usr/bin/env python3
|
|
"""
|
|
Autonomous Coding Agent Demo
|
|
============================
|
|
|
|
A minimal harness demonstrating long-running autonomous coding with Claude.
|
|
This script implements a unified orchestrator pattern that handles:
|
|
- Initialization (creating features from app_spec)
|
|
- Coding agents (implementing features)
|
|
- Testing agents (regression testing)
|
|
|
|
Example Usage:
|
|
# Using absolute path directly
|
|
python autonomous_agent_demo.py --project-dir C:/Projects/my-app
|
|
|
|
# Using registered project name (looked up from registry)
|
|
python autonomous_agent_demo.py --project-dir my-app
|
|
|
|
# Limit iterations for testing (when running as subprocess)
|
|
python autonomous_agent_demo.py --project-dir my-app --max-iterations 5
|
|
|
|
# YOLO mode: rapid prototyping without testing agents
|
|
python autonomous_agent_demo.py --project-dir my-app --yolo
|
|
|
|
# Parallel execution with 3 concurrent coding agents
|
|
python autonomous_agent_demo.py --project-dir my-app --concurrency 3
|
|
|
|
# Single agent mode (orchestrator with concurrency=1, the default)
|
|
python autonomous_agent_demo.py --project-dir my-app
|
|
|
|
# Run as specific agent type (used by orchestrator to spawn subprocesses)
|
|
python autonomous_agent_demo.py --project-dir my-app --agent-type initializer
|
|
python autonomous_agent_demo.py --project-dir my-app --agent-type coding --feature-id 42
|
|
python autonomous_agent_demo.py --project-dir my-app --agent-type testing
|
|
"""
|
|
|
|
import argparse
|
|
import asyncio
|
|
from pathlib import Path
|
|
|
|
from dotenv import load_dotenv
|
|
|
|
# Load environment variables from .env file (if it exists)
|
|
# IMPORTANT: Must be called BEFORE importing other modules that read env vars at load time
|
|
load_dotenv()
|
|
|
|
from agent import run_autonomous_agent
|
|
from registry import DEFAULT_MODEL, get_project_path
|
|
|
|
|
|
def parse_args() -> argparse.Namespace:
|
|
"""Parse command line arguments."""
|
|
parser = argparse.ArgumentParser(
|
|
description="Autonomous Coding Agent Demo - Unified orchestrator pattern",
|
|
formatter_class=argparse.RawDescriptionHelpFormatter,
|
|
epilog="""
|
|
Examples:
|
|
# Use absolute path directly (single agent, default)
|
|
python autonomous_agent_demo.py --project-dir C:/Projects/my-app
|
|
|
|
# Use registered project name (looked up from registry)
|
|
python autonomous_agent_demo.py --project-dir my-app
|
|
|
|
# Parallel execution with 3 concurrent agents
|
|
python autonomous_agent_demo.py --project-dir my-app --concurrency 3
|
|
|
|
# YOLO mode: rapid prototyping without testing agents
|
|
python autonomous_agent_demo.py --project-dir my-app --yolo
|
|
|
|
# Configure testing agent ratio (2 testing agents per coding agent)
|
|
python autonomous_agent_demo.py --project-dir my-app --testing-ratio 2
|
|
|
|
# Disable testing agents (similar to YOLO but with verification)
|
|
python autonomous_agent_demo.py --project-dir my-app --testing-ratio 0
|
|
|
|
Authentication:
|
|
Uses Claude CLI authentication (run 'claude login' if not logged in)
|
|
Authentication is handled by start.bat/start.sh before this runs
|
|
""",
|
|
)
|
|
|
|
parser.add_argument(
|
|
"--project-dir",
|
|
type=str,
|
|
required=True,
|
|
help="Project directory path (absolute) or registered project name",
|
|
)
|
|
|
|
parser.add_argument(
|
|
"--max-iterations",
|
|
type=int,
|
|
default=None,
|
|
help="Maximum number of agent iterations (default: unlimited, typically 1 for subprocesses)",
|
|
)
|
|
|
|
parser.add_argument(
|
|
"--model",
|
|
type=str,
|
|
default=DEFAULT_MODEL,
|
|
help=f"Claude model to use (default: {DEFAULT_MODEL})",
|
|
)
|
|
|
|
parser.add_argument(
|
|
"--yolo",
|
|
action="store_true",
|
|
default=False,
|
|
help="Enable YOLO mode: skip testing agents for rapid prototyping",
|
|
)
|
|
|
|
# Unified orchestrator mode (replaces --parallel)
|
|
parser.add_argument(
|
|
"--concurrency", "-c",
|
|
type=int,
|
|
default=1,
|
|
help="Number of concurrent coding agents (default: 1, max: 5)",
|
|
)
|
|
|
|
# Backward compatibility: --parallel is deprecated alias for --concurrency
|
|
parser.add_argument(
|
|
"--parallel", "-p",
|
|
type=int,
|
|
nargs="?",
|
|
const=3,
|
|
default=None,
|
|
metavar="N",
|
|
help="DEPRECATED: Use --concurrency instead. Alias for --concurrency.",
|
|
)
|
|
|
|
parser.add_argument(
|
|
"--feature-id",
|
|
type=int,
|
|
default=None,
|
|
help="Work on a specific feature ID only (used by orchestrator for coding agents)",
|
|
)
|
|
|
|
# Agent type for subprocess mode
|
|
parser.add_argument(
|
|
"--agent-type",
|
|
choices=["initializer", "coding", "testing"],
|
|
default=None,
|
|
help="Agent type (used by orchestrator to spawn specialized subprocesses)",
|
|
)
|
|
|
|
parser.add_argument(
|
|
"--testing-feature-id",
|
|
type=int,
|
|
default=None,
|
|
help="Feature ID to regression test (used by orchestrator for testing agents, legacy single mode)",
|
|
)
|
|
|
|
parser.add_argument(
|
|
"--testing-feature-ids",
|
|
type=str,
|
|
default=None,
|
|
help="Comma-separated feature IDs to regression test in batch (e.g., '5,12,18')",
|
|
)
|
|
|
|
# Testing agent configuration
|
|
parser.add_argument(
|
|
"--testing-ratio",
|
|
type=int,
|
|
default=1,
|
|
help="Testing agents per coding agent (0-3, default: 1). Set to 0 to disable testing agents.",
|
|
)
|
|
|
|
parser.add_argument(
|
|
"--testing-batch-size",
|
|
type=int,
|
|
default=3,
|
|
help="Number of features per testing batch (1-5, default: 3)",
|
|
)
|
|
|
|
return parser.parse_args()
|
|
|
|
|
|
def main() -> None:
|
|
"""Main entry point."""
|
|
print("[ENTRY] autonomous_agent_demo.py starting...", flush=True)
|
|
args = parse_args()
|
|
|
|
# Note: Authentication is handled by start.bat/start.sh before this script runs.
|
|
# The Claude SDK auto-detects credentials from ~/.claude/.credentials.json
|
|
|
|
# Handle deprecated --parallel flag
|
|
if args.parallel is not None:
|
|
print("WARNING: --parallel is deprecated. Use --concurrency instead.", flush=True)
|
|
args.concurrency = args.parallel
|
|
|
|
# Resolve project directory:
|
|
# 1. If absolute path, use as-is
|
|
# 2. Otherwise, look up from registry by name
|
|
project_dir_input = args.project_dir
|
|
project_dir = Path(project_dir_input)
|
|
|
|
if project_dir.is_absolute():
|
|
# Absolute path provided - use directly
|
|
if not project_dir.exists():
|
|
print(f"Error: Project directory does not exist: {project_dir}")
|
|
return
|
|
else:
|
|
# Treat as a project name - look up from registry
|
|
registered_path = get_project_path(project_dir_input)
|
|
if registered_path:
|
|
project_dir = registered_path
|
|
else:
|
|
print(f"Error: Project '{project_dir_input}' not found in registry")
|
|
print("Use an absolute path or register the project first.")
|
|
return
|
|
|
|
# Migrate project layout to .autocoder/ if needed (idempotent, safe)
|
|
from autocoder_paths import migrate_project_layout
|
|
migrated = migrate_project_layout(project_dir)
|
|
if migrated:
|
|
print(f"Migrated project files to .autocoder/: {', '.join(migrated)}", flush=True)
|
|
|
|
# Parse batch testing feature IDs (comma-separated string -> list[int])
|
|
testing_feature_ids: list[int] | None = None
|
|
if args.testing_feature_ids:
|
|
try:
|
|
testing_feature_ids = [int(x.strip()) for x in args.testing_feature_ids.split(",") if x.strip()]
|
|
except ValueError:
|
|
print(f"Error: --testing-feature-ids must be comma-separated integers, got: {args.testing_feature_ids}")
|
|
return
|
|
|
|
try:
|
|
if args.agent_type:
|
|
# Subprocess mode - spawned by orchestrator for a specific role
|
|
asyncio.run(
|
|
run_autonomous_agent(
|
|
project_dir=project_dir,
|
|
model=args.model,
|
|
max_iterations=args.max_iterations or 1,
|
|
yolo_mode=args.yolo,
|
|
feature_id=args.feature_id,
|
|
agent_type=args.agent_type,
|
|
testing_feature_id=args.testing_feature_id,
|
|
testing_feature_ids=testing_feature_ids,
|
|
)
|
|
)
|
|
else:
|
|
# Entry point mode - always use unified orchestrator
|
|
from parallel_orchestrator import run_parallel_orchestrator
|
|
|
|
# Clamp concurrency to valid range (1-5)
|
|
concurrency = max(1, min(args.concurrency, 5))
|
|
if concurrency != args.concurrency:
|
|
print(f"Clamping concurrency to valid range: {concurrency}", flush=True)
|
|
|
|
asyncio.run(
|
|
run_parallel_orchestrator(
|
|
project_dir=project_dir,
|
|
max_concurrency=concurrency,
|
|
model=args.model,
|
|
yolo_mode=args.yolo,
|
|
testing_agent_ratio=args.testing_ratio,
|
|
testing_batch_size=args.testing_batch_size,
|
|
)
|
|
)
|
|
except KeyboardInterrupt:
|
|
print("\n\nInterrupted by user")
|
|
print("To resume, run the same command again")
|
|
except Exception as e:
|
|
print(f"\nFatal error: {e}")
|
|
raise
|
|
|
|
|
|
if __name__ == "__main__":
|
|
main()
|