jeffy-loop 1.23.0

Quickstart  ·  Usage  ·  How it works  ·  The receipts  ·  Headless  ·  White paper

Autonomous Engineering With Proof

Agents that don’t just act.
They audit · verify · attack · and prove.

Jeffy Loop is an autonomous engineering system built around a simple principle: AI agents shouldn’t just produce work. They should produce evidence that the work is correct. Instead of asking an agent to complete a task and trusting its conclusion, Jeffy creates a continuous Audit → Attack → Verify → Prove loop in which specialized agents inspect the work, challenge it, validate the result, and generate an auditable record of what happened. The goal isn’t simply autonomous code generation; it is autonomous engineering with proof - where every claimed result is accompanied by reproducible evidence that can be independently examined.

Jeffy treats “done” as something that must be demonstrated, not declared. The system is designed to turn autonomous engineering from a conversational interaction into an evidence-producing process: actions leave traces, decisions have provenance, failures are exposed rather than hidden, and successful outcomes produce a durable receipt of what was changed, why it was changed, and how the result was verified.

The proof

Jeffy was run against 132 open-source projects with no connection to this repository, each judged by its own test suite, every run published, failures included.

_{103 projects run to convergence across 13 languages with no language-specific analyzer or ruleset. Derived from the scorecard at render time.}

Independent Validation

A merged pull request is the one result Jeffy cannot award itself. It takes an independent maintainer, someone with no stake in this project, to review the patch and accept it into their own codebase. Maintainers across 21 open source projects have done exactly that, including:

• Google, snappy - every release build compressed a 4 GiB input into a stream whose header claimed 0 bytes; merged the morning after filing.
• Apple (2)
  • swift-log - a documented no-op setter asserted instead; merged after the maintainer asked for the doc-only form.
  • swift-protobuf - the project's own CMake build of protoc-gen-swift had not compiled since June; approved and merged the day it was filed.
• Microsoft, mimalloc - the zeroing allocator returned uninitialized memory above the small-size threshold; merged by the author the same day.
• Apache (2)
  • commons-text - a StringMatcher overload forwarded the buffer end as its start; merged the morning after review.
  • commons-csv - the record counter's Javadoc said headers were not counted while the constructor's header was; corrected, merged three days after filing.
• JetBrains, kotlinx-datetime - deprecation quick-fixes pointed developers at the wrong replacement; merged within two hours.
• Node.js, ada - the URL parser reported host_end one byte short; merged twelve minutes after filing.

Every project, every patch, and every failure

Quickstart

You need Claude Code, signed in once, and git. The installer handles everything else, including jq.

git clone https://github.com/lenamonj/jeffy-loop.git
cd jeffy-loop
./install.sh        # Windows PowerShell: .\install.ps1

Or, with uv, no clone needed:

uv tool install jeffy-loop
jeffy install

[!NOTE] If PowerShell refuses with "running scripts is disabled on this system", run powershell -ExecutionPolicy Bypass -File .\install.ps1 once.

Open Claude Code in the project you want to improve and type /jeffy 10. It is a slash command inside the session, not a shell command. If you only want the Highs found and fixed, type /jeffy 10 --highs; it stops at the first audit that finds none, so it is usually the faster run (High hunt). When the run ends, start a new session to run it again; the restart is doing real work.

Five guarantees

Each one is enforced by the iteration prompt, the state files, or the Stop hook. How.

1. It audits like an engineer, not a linter. A finding exists only when the loop can point at it and prove it with a runnable check.
2. It cannot wreck your repo. Every iteration is a local commit, a broken verify is reverted, and nothing is ever pushed.
3. "Done" is not the agent's opinion. An adversarial evaluator and a shell gate re-check every declaration.
4. It cannot claim what it never looked at. The whole public surface goes on a checklist before any finding is filed.
5. Lessons become machinery. A rule learned once binds every later iteration, and the engine itself passes 347 behavioural checks on Linux, Windows and macOS.

Documentation

[!IMPORTANT] Trust model. The engine is one shell script, skills/jeffy/hooks/stop-hook.sh, plus the small library beside it in skills/jeffy/hooks/lib/, registered as a Claude Code Stop hook. In a session with no live Jeffy state file it exits at once and does nothing. The installer writes two skill folders under ~/.claude/skills, one hook entry in ~/.claude/settings.json, and, only if you say yes when jq is missing, a jq install through your package manager.

License

MIT