Menu
Task: {{ task }} Final best success_rate: {{ best_score }} (target was {{ target_success }}). Write the markdown report for this hill-climb run.
Install with Codex or Claude Copy this prompt, paste it into Codex, Claude, or another assistant, and let it review the skill page and install it for you.
Based on SOC occupation classification
Run a tiny human-in-the-loop session: ask the user a couple of questions on the console, then write a short personalized note from their answers.
Task: {{ task }} Implement (or, if a proposal is given below, modify) the ML pipeline so it trains and evaluates end-to-end. Target metric to beat: {{ target_accuracy }}. Proposed change for THIS experiment (empty on the baseline): {{ current_proposal | default("(none — build a simple baseline)") }}
Task: {{ task }} Best VALIDATION accuracy (the hill-climb selection metric): {{ best_score }} (target {{ target_accuracy }}, higher is better). Held-out TEST accuracy of the retrained winner — the HEADLINE number, selected on validation and reported once on the test set: {{ final_test_score }}. Write the final HTML research report for this ML auto-research run.
Review the applied change against the proposal, check the contract, and run the smoke tests. Decide pass or fail. Proposal that was supposed to be applied: {{ current_proposal | default("(none — baseline build)") }}
Competition: {{ competition_id }} Metric: {{ metric_name }} ({{ "lower is better" if lower_is_better else "higher is better" }}). Final best validation score: {{ best_score }} (target {{ target_score }}). Write the final HTML report for this kaggle-solver run.
Condense the current kaggle experiment proposal into one short paragraph for the running research log.
| name | report_narrative |
| description | Task: {{ task }} Final best success_rate: {{ best_score }} (target was {{ target_success }}). Write the markdown report for this hill-climb run. |
| tools | ["read_file","write_file","run_command","git_log"] |
SKILL_ID: report_narrative
You are writing the final report for an automated hill-climb over LeWM training on OGBench-Cube. First gather the facts:
research_log.md — the goal, the paper's reference numbers, and one line per
experiment (keep = improved and was committed, revert = did not improve).experiments.jsonl — structured ledger: each record has the candidate
score, the running best, status, and the proposal text that produced it.git_log — the kept experiments are commits on this branch
(saage: keep experiment, ...).config/train/lewm.yaml / config/train/model/lewm.yaml — the
winning configuration that survived.Then write report_narrative.md in markdown. Be ACCURATE — use only the real
scores and experiments from those files; do not invent results. Include:
Baseline score at the fixed {{ train_epochs }}-epoch budget, the final best score, and how both compare to the paper's 74% target and the earlier manual runs (60-64% at epoch 54 of a 100-epoch run).
The KEPT experiments in order: what changed, why it plausibly helped, score before -> after. Then notable reverted ideas and why a fixed short budget might reject them.
2-4 bullets — e.g. which surviving config to scale up to a full-length training run, and what to try next if the target was not reached.
Write ONLY to report_narrative.md. Finish with a one-line confirmation.