i open-sourced loop rat - an autonomous loop agent for Claude
the rule it never breaks: whoever graded the work is not who did it
> https://t.co/3ikLxMDhal
check the receipts and you'll see exactly which loop got stopped by the guard, how much of the day's budget is left, and which shift is still sitting there unread
nothing in those receipts is a summary. it's the actual trace, phase by phase
passing tests isn't proof by itself - an exit code is
i'm fully endorsing the Loop Rat token. to show it, the CA goes in my bio
CA: 0x642d30c84211ade7768fe557fbaed7224e2068c7
LOOP RAT: THE TECH UNDER AN AUTONOMOUS CODING LOOP
no framework, no service, no database
a dozen files of bash and python, plus one https://t.co/VUDfqGE1Fp per loop
one shift runs six phases:
> preflight - won't start if state/HALT exists, spend is gone, or the loop's already running
> act - runs https://t.co/VUDfqGE1Fp under a hard timeout. shell gathers facts, the model only judges
> verify - runs the plan's verify: command. "tests pass" isn't proof. the exit code is
> guard - fingerprints the tree. blocks a denylisted path, an oversized diff, anything shaped like a secret
> grade - a second agent, fresh context, grades against rubrics it never wrote
> receipt - writes receipt.json, appends the trace, updates the checkpoint
three brakes stack in order:
> daily spend cap, checked before the shift starts
> per-call cap, checked before each model call
> hard timeout for whatever's already running
scheduling runs on schedule.yml and one cron line
run-due is idempotent - a missed window is a skipped shift, never a backlog dumped at once
The official ca: 0x642d30c84211ade7768fe557fbaed7224e2068c7
50 tests cover the parsers, the lock, the guard, the ledger under concurrent writers, and the kill switch
https://t.co/3ikLxMDhal