Configuration management for autonomous coding. Use when loading settings, managing environment variables, configuring providers, or setting up autonomous mode options.
/plugin marketplace add adaptationio/Skrillz/plugin install skrillz@skrillzThis skill inherits all available tools. When active, it can use any tool Claude has access to.
scripts/config_manager.pyCentralized configuration management for the Auto-Claude replication skill set.
Manages all configuration for autonomous coding:
from scripts.config_manager import ConfigManager
config = ConfigManager(project_dir)
settings = config.load()
# Access settings
model = settings.model # claude-opus-4-5-20251101
max_iterations = settings.max_iterations # 50
config.create_default_config()
# Creates .claude/autonomous-config.json with defaults
config.update({
"max_iterations": 30,
"max_cost_usd": 15.00,
"verbose": True
})
{
"enabled": true,
"objective": "Build feature X",
"success_criteria": [
"All tests pass",
"Code coverage > 80%"
],
"max_iterations": 50,
"max_cost_usd": 20.00,
"max_consecutive_failures": 3,
"max_runtime_minutes": 480,
"analyzer_model": "claude-sonnet-4-20250514",
"verbose": false,
"notify_on_complete": true
}
| Variable | Required | Default | Description |
|---|---|---|---|
ANTHROPIC_API_KEY | Yes | - | API key for Claude |
AUTO_BUILD_MODEL | No | claude-opus-4-5-20251101 | Model for builds |
DEFAULT_BRANCH | No | auto-detect | Base branch |
GRAPHITI_ENABLED | No | true | Enable memory |
GRAPHITI_LLM_PROVIDER | No | openai | LLM for memory |
GRAPHITI_EMBEDDER_PROVIDER | No | openai | Embeddings |
DEBUG | No | false | Debug logging |
DEBUG_LEVEL | No | 1 | Verbosity (1-3) |
@dataclass
class BuildConfig:
model: str = "claude-opus-4-5-20251101"
max_thinking_tokens: dict = field(default_factory=lambda: {
'planner': 5000,
'coder': None,
'qa_reviewer': 10000,
'qa_fixer': None
})
max_iterations: int = 50
max_parallel_agents: int = 4
skip_qa: bool = False
timeout_ms: int = 600000
settings = config.load()
# Merges: defaults → env vars → config file → overrides
errors = config.validate()
if errors:
for error in errors:
print(f"Config error: {error}")
memory_config = config.get_memory_config()
# Returns Graphiti configuration
build_config = config.get_build_config()
# Returns build settings
paths = config.get_paths()
# Returns:
# specs_dir: .auto-claude/specs/
# worktrees_dir: .worktrees/auto-claude/
# memory_dir: .claude/memory/
# checkpoints_dir: .claude/checkpoints/
| Setting | Default | Description |
|---|---|---|
enabled | false | Autonomous mode |
max_iterations | 50 | Max loop iterations |
max_cost_usd | 20.00 | Budget limit |
max_consecutive_failures | 3 | Before escalation |
max_runtime_minutes | 480 | 8 hour limit |
context_threshold | 0.85 | Trigger handoff |
auto_checkpoint | true | Create checkpoints |
references/CONFIG-SCHEMA.md - Full schema documentationreferences/ENVIRONMENT.md - Environment variable guidescripts/config_manager.py - Core ConfigManager classscripts/config_schema.py - Configuration schemasscripts/path_manager.py - Path utilitiesActivates when the user asks about Agent Skills, wants to find reusable AI capabilities, needs to install skills, or mentions skills for Claude. Use for discovering, retrieving, and installing skills.
Activates when the user asks about AI prompts, needs prompt templates, wants to search for prompts, or mentions prompts.chat. Use for discovering, retrieving, and improving prompts.
This skill should be used when the user asks to "create an agent", "add an agent", "write a subagent", "agent frontmatter", "when to use description", "agent examples", "agent tools", "agent colors", "autonomous agent", or needs guidance on agent structure, system prompts, triggering conditions, or agent development best practices for Claude Code plugins.