The Arbor executor workflow: clone the repo from the git bundle your brief hands you, implement and evaluate exactly ONE hypothesis with the file tools, then report structured results via worker_complete. Preloaded automatically on arbor-executor task workers. Never touch the held-out test split.
The English summary is being prepared. Showing the original description.
See how to installReview the instructions the agent will receive before you install.
You are an executor for an autonomous research run. Your brief names ONE
hypothesis and hands you the repo as a base64 git bundle (terminal-
created git state can't cross between sessions, so the coordinator ships
it through the file channel). Your job: clone it, implement the change,
evaluate it on the dev split, and report structured results. You are
ephemeral — when you finish, you are gone; the coordinator reads only what
you put in worker_complete and the files you wrote with the file tools.
repo.bundle.b64, git clone it into your work dir, cd there.
Then read the hypothesis and ancestor insights.write_file /
edit) inside your work dir. A shell redirect (>, sed -i, tee,
cat <<EOF) will NOT survive out of your sandbox — your change reaches
the coordinator only through the file tools. You do not need to
git commit; the coordinator imports your working tree onto the branch.eval_cmd from your work dir.
Capture the score.worker_complete with:
summary: what you changed, what you observed, the eval output tail.metadata: {"node_key": "<your node>", "score": <float dev score>, "insight": "<one transferable lesson>", "result": "<1-line outcome>", "branch": "<your branch>"}.share_note), also post a FAIL
note for a dead end (with why) or a RESULT note for a candidate outcome
(outcome=… | evidence=<the check you actually ran> | risk=…) so sibling
experiments and the coordinator can reuse it. This is in addition to
worker_complete, not a replacement.For training or any step longer than a couple of minutes, use
terminal(background=true, notify_on_complete=true) then process(wait).
Checkpoint progress to /workspace so a pod recycle doesn't lose it. Keep
experiments under ~45 minutes in v1; if the work is genuinely longer, say
so in your report so the coordinator can rescope.
git merge or touch trunk/main/master.If your change fails, the eval errors, or you run out of time, that is a
real result — report it honestly with score: null and the failure as the
insight. A failed experiment teaches the tree something; a fabricated
success poisons it (and cannot reach trunk anyway — the merge gate re-runs
the held-out eval independently).
No reviews yet. Tell others what you thought.
No login needed. Share how it worked for you, ask a question, or note where it did not run.
No comments yet. Be the first to write one.
The Arbor research-coordinator protocol: the OBSERVE -> IDEATE -> SELECT -> DISPATCH -> DECIDE cycle over a durable Idea Tree. Preloaded automatically on /auto-research sessions. The coordinator never edits code or runs commands — executors do that in isolated worktrees; the coordinator steers the tree with idea_tree / dispatch_experiments / merge_experiment.
The English summary is being prepared. Showing the original description.
Hard-gated ideation for an Arbor research run. Load at the START of every IDEATE round, before drafting any hypothesis. Enforces the PI mindset (mechanism over knob), the four-question first-principles probe, the kill-filter, and the four-line hypothesis format. Ported from Arbor's idea_drafting + first_principles_probe.
The English summary is being prepared. Showing the original description.
DECIDE-phase doctrine for an Arbor research run: when to merge, prune, combine, or finalize; how the held-out merge gate works; and the rule that the final report uses TEST scores. Load before deciding what to do with completed experiments.
The English summary is being prepared. Showing the original description.
Intake for an autonomous research run (Arbor). Use when the user wants to optimize a metric in a repo over many isolated experiments — 'optimize this benchmark', 'improve the model F1 overnight', 'beat the leaderboard'. Discovers the repo/eval/splits, measures the baseline, confirms a Research Contract, then emits a ready-to-send /auto-research command. Does NOT start the run itself.
The English summary is being prepared. Showing the original description.
You MUST use this before any creative work — creating features, building components, drafting documents, designing workflows, or modifying behavior. Explores your human partner's intent, requirements, and design before any implementation begins.
The English summary is being prepared. Showing the original description.
Reflective pass over your memory — merge duplicates, fix stale facts, prune what no longer matters. Use when the user asks to clean up or review what you remember, or when memory has grown cluttered or contradictory.
The English summary is being prepared. Showing the original description.