Orchestrates the full software development lifecycle through autonomous task execution, parallel research agents, and structured validation gates.
The lev-sdlc skill provides a comprehensive command surface for managing the Leviathan SDLC pipeline within Claude Code. It enables developers to move entities through a structured state machine—from initial draft to validated code—using specialized AI model profiles for execution, research, and hygiene scans. By fanning out complex plans into parallel sub-agents and enforcing strict validation gates, this skill ensures that autonomous development remains high-quality, documented, and aligned with project-specific BDD contracts.
主要功能
01Parallel plan decomposition for deep-research sub-agent orchestration
022 GitHub stars
03Structured entity lifecycle tracking from draft to validated production code
04Autonomous development loops with heartbeat scheduling and budget controls
05Append-only instruction logging for persistent project rules and learnings
06Multistage task execution using specialized Opus and Sonnet model profiles
使用场景
01Decomposing complex epics into parallel research and implementation sub-tasks
02Managing a human-in-the-loop review queue for autonomous agent contributions
03Running periodic project hygiene scans to align documentation with codebases
