Can I set a maximum spend for autonomous tasks?

Yes, the skill supports a --budget flag that allows you to set a hard token cap, ensuring the agent stops immediately once the limit is reached.

What is the benefit of in-process execution vs. cron jobs?

In-process execution preserves state continuity, meaning the agent doesn't have to re-load the entire project context at every interval, leading to faster and more coherent development loops.

What happens if the agent encounters repeated errors?

Autodev Lev features a built-in circuit breaker that triggers a shutdown after three consecutive unsuccessful ticks to prevent wasting resources on structural issues.

How does Autodev Lev save on token costs?

It utilizes a local, zero-cost 'Scan' phase to identify and prioritize tasks before engaging the LLM, ensuring that tokens are only spent on actual execution and validation.

Autonomous SDLC Heartbeat

Name: Autonomous SDLC Heartbeat
Author: lev-os

bylev-os

•

Productivity & Workflow

Executes a persistent, heartbeat-driven autonomous development loop with advanced state continuity and token budget management.

Autonomous SDLC Heartbeat (autodev-lev) is a specialized skill designed for deep, multi-stage development tasks. Unlike standard stateless automations, this skill runs an in-process loop that preserves context and state between 'heartbeats,' significantly improving token efficiency. It moves through a structured lifecycle of scanning, planning, execution, and validation, allowing Claude to autonomously manage complex feature implementations or codebase refactors while respecting strict user-defined budget and safety constraints.

Key Features

01Zero-cost scan phase to prioritize entities without consuming tokens

02Configurable safety controls including token budgets and tick limits

03Heartbeat-driven execution for stateful persistence between tasks

04Smart circuit breakers to prevent runaway spend or repetitive failures

05Multi-stage SDLC orchestration (Scan, Plan, Exec, Validate)

062 GitHub stars

Use Cases

01Long-running development tasks requiring state continuity across multiple iterations

02Continuous background codebase maintenance and hygiene improvement

03Autonomous execution of multi-step feature roadmaps and complex refactors

Key Features

01Zero-cost scan phase to prioritize entities without consuming tokens

02Configurable safety controls including token budgets and tick limits

03Heartbeat-driven execution for stateful persistence between tasks

04Smart circuit breakers to prevent runaway spend or repetitive failures

05Multi-stage SDLC orchestration (Scan, Plan, Exec, Validate)

062 GitHub stars

Use Cases

01Long-running development tasks requiring state continuity across multiple iterations

02Continuous background codebase maintenance and hygiene improvement

03Autonomous execution of multi-step feature roadmaps and complex refactors