"author:anksarvesh-lgtm" matching MCP servers:

Enables supervised, crash-resilient autonomous AI pair programming through nine MCP tools that checkpoint task progress in an atomic local DAG ledger, keep the machine awake during long runs, summarize repositories within a token budget, count tokens and costs across models, and rotate accounts before quota runs out. All state stays in a sandboxed project-local directory with zero telemetry, so long multi-step agent sessions survive crashes, context limits, and premature stopping.