name: codex-autoresearch description: "Run repeated, measured Git experiments toward a numeric target; keep improvements and revert failures. Use for autonomous optimization or managing an autoresearch run, not one-shot edits." metadata: short-description: "Run measurable autonomous experiments"

Codex Autoresearch

Improve a repository through repeated, reversible experiments:

hypothesize -> change -> measure -> learn -> keep or revert -> repeat

Codex chooses hypotheses and makes code changes. The control script owns measurement, commits, rollback, and the event history.

Respond To The Request

Resolve <control> to this skill's own scripts/autoresearch.py; do not assume it is installed in the target repository.

For a status or results request, run the corresponding command directly:

Report the requested result without starting or resuming experiments. Return the generated path for an HTML report.

Before starting or resuming an experiment, read the applicable workflow unless it is already in context:

• Workflow: starting, continuing, resuming, or changing an experiment.
• Background: launching or controlling a detached run.

Experiment Boundary

• One run owns one Git repository, one numeric metric, one confirmed target, and approved path scopes.
• For a new foreground run, initialize successfully before creating its Goal; the Goal identifies the returned run id. If initialization reports complete, no Goal is needed.
• Use finish to finalize each coherent experiment. Do not manually commit, revert, or edit run artifacts during an active run.
• Only a verified target can mean complete; an iteration limit, error, or external blocker has its own status.
• Validate state through the control script on entry or resume. Use recorded experiments rather than guessing from conversation.
• Surface command and state errors with their diagnostic paths. Never bypass validation, fabricate metrics, or reconstruct missing events.

Foreground continuation belongs to the official Codex Goal. Background continuation belongs to the detached controller. An already launched background worker follows its supplied experiment contract, without starting a new launch flow.

Tell Codex what measurable result you want. Codex inspects the repository, confirms the experiment with you, changes one thing, verifies it, keeps improvements, reverts failures, and repeats until the target is reached.

Autoresearch works for test failures, coverage, type errors, warnings, latency, binary size, reproducible security findings, and any other outcome a command can measure.

Quick Start

Install from Codex:

$skill-installer install https://github.com/leo-lilinxiao/codex-autoresearch

Open a clean Git repository with Full Access:

codex --dangerously-bypass-approvals-and-sandbox

Then invoke the skill:

You:   $codex-autoresearch
       Reduce `python3 scripts/score.py` error_count to 0.

Codex: Baseline: 5
       Target: 0 (lower is better)
       Scope: src/
       Verify: python3 scripts/score.py, JSON key error_count
       Guard: none
       Run in foreground or background?

You:   Background. Go.

Codex launches the confirmed run. No Codex configuration changes or special prompt syntax are required.

See Installation for manual and development installs.

The Loop

inspect evidence
      |
change one focused thing
      |
commit and measure
      |
      +-- improved + guard passes --> keep
      |
      +-- otherwise ---------------> revert
      |
append an audit event
      |
repeat until target

The control script owns commits, verification, rollback, and state. Codex owns the hypotheses and code changes.

Foreground And Background

Foreground and background use the same experiment rules. A run uses one mode at a time. Foreground continuation uses a Codex Goal; background continuation belongs to the detached controller.

What Gets Confirmed

Before the first write, Codex shows:

• the goal and numeric target;
• repository-relative paths it may change;
• the metric command and explicit parser;
• an optional regression guard;
• foreground or background mode;
• an optional iteration limit.

Initialization requires a clean named Git branch. One run manages one repository.

Codex reuses choices and approval you have already given. Once the experiment is confirmed, it keeps working without asking you to approve each iteration.

Results

Run artifacts live in autoresearch-results/ and stay uncommitted:

events.jsonl is the state history. Missing, malformed, contradictory, or partial state is an error; the skill never guesses a result from old files or conversational memory.

Review Results

Ask the skill to show the validated experiment history:

$codex-autoresearch show experiment history

Codex Autoresearch
Run: 0a516883  Status: complete  Mode: foreground
Metric: error_count  2 -> 0  Target: 0 (lower is better)

SEQ  ITER  EVENT     PREVIOUS  TRIAL  RETAINED  DESCRIPTION
---  ----  --------  --------  -----  --------  ------------------------------------
  0     0  baseline         -      -         2  Initial measurement
  1     1  discard          2      3         2  Broaden parser fallback
  2     2  keep             2      1         1  Fix nested parser branch
  3     3  keep             1      0         0  Remove final parser error
  4     3  complete         -      -         0  retained metric satisfies the target

The same validated events can be exported as TSV or rendered as a self-contained static report:

$codex-autoresearch export experiment history as TSV
$codex-autoresearch generate an HTML report

The report is written to autoresearch-results/report.html. It is a replaceable snapshot, not runtime state.

Safety Model

• Every trial is a Git commit.
• A non-improving trial or failed guard is reverted with git revert.
• Out-of-scope edits, branch changes, HEAD drift, malformed metrics, command failures, timeouts, and generated byproducts stop the run with an exact error and log path.
• Autoresearch artifacts are never staged.
• A run reports complete only when the retained metric reaches the confirmed target.
• A genuine external blocker is reported explicitly; a difficult or unsuccessful hypothesis is not treated as blocked.

This strictness is intentional. Silent recovery makes long autonomous runs impossible to trust.

Good Metrics

The verify command must exit successfully and place one finite number on its final non-empty stdout line. It may instead print a JSON object on that line when Codex names one numeric key explicitly.

7

{"error_count": 7, "passed": 12}

Use a guard for behavior the metric does not protect, such as a test suite around a latency benchmark. The guard must pass at baseline.

Documentation

FAQ

Does installation change my Codex settings?

No. Installation copies the skill files. Use a current Codex release so foreground runs can use the built-in Goal capability.

Why Full Access?

Each iteration creates or reverts a Git commit. Restricted sandboxes may block writes under .git. Background runs therefore default to Full Access; workspace-write remains an explicit option when its limitations are acceptable.

Can I stop and resume?

Yes. Interrupt or pause a foreground Goal. For background, invoke $codex-autoresearch and ask for status, stop, or resume. A new direction is optional; "resume" continues the confirmed experiment.

Can it run without Git or across several repos?

No. Git is the experiment memory and rollback boundary. Use one run per repository so commit ownership and metrics remain unambiguous.

Is this only for small changes?

No. One experiment should test one coherent hypothesis. Its size should match the hypothesis, while still being independently measurable and reversible.

Acknowledgments

Inspired by Karpathy's autoresearch, generalized for Codex and software repositories.

Citation

@misc{codex-autoresearch,
  author = {Li, Linxiao},
  title = {Codex Autoresearch: Autonomous Goal-Driven Experimentation for Codex},
  year = {2026},
  publisher = {GitHub},
  url = {https://github.com/leo-lilinxiao/codex-autoresearch}
}

GitHub also reads CITATION.cff for its Cite this repository menu.

Star History

License

MIT, see LICENSE.