菜单
Use when running a new experiment. Follows the two-phase protocol from LAB.md.
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
Use periodically (weekly or before a release) to find stale numbers, outdated descriptions, broken links, and inconsistencies across the codebase. Run after merging multiple PRs or before preparing a meeting report.
Use before a Sutro Group meeting to compile results and prepare a presentation report.
Use before any research task, experiment, or PR review. Loads current project state from DISCOVERIES.md, open questions, and recent Telegram discussion.
Use at the start of any session and before pushing. Syncs Telegram, Google Docs, and GitHub state for the Sutro Group research workspace.
Use at the start of a weekly session. Syncs all sources and generates a catch-up summary.
Use when drafting, editing, or reviewing any prose to detect and remove AI writing patterns including overused vocabulary (delve, tapestry, landscape), formulaic structures (binary contrasts, rule of three), throat-clearing openers, business jargon, and other LLM tells
| name | run-experiment |
| description | Use when running a new experiment. Follows the two-phase protocol from LAB.md. |
New method: not in the registry (search_space.yaml). Create a new experiment file from the template. Add the method to the registry if it works.
Existing method, new config: method is in the registry but you're testing a different configuration (different n, k, hyperparameters). Use the existing experiment code or copy and modify.
Either way, the steps and output format are the same.
Read DISCOVERIES.md. Check what's already proven. Do not repeat existing experiments.
Identify the hypothesis. Either from TODO.md, research/questions.yaml, or the user's request. State it as: "If we do X, then Y will happen because Z."
Create the experiment file. Copy src/sparse_parity/experiments/_template.py. Change one variable from the baseline.
Run the experiment. Capture results including accuracy, ARD, DMC, wall time. Record seed, config, environment (Python version, numpy version, OS, git hash).
Save Phase 1 output. Write results/{exp_id}/results.json with raw numbers, config, and environment. No interpretation in this file.
Verify. Re-run with a different seed. If the result only holds on one seed, note that.
Write Phase 2 findings. Create docs/findings/{exp_id}.md using the template from LAB.md. Use Status: SUCCESS | PARTIAL | FAILED (not "COMPLETED"). Reference the results JSON. Add analysis and impact.
Classify in research/log.jsonl. Use "class": "WIN" only if the result is a clear improvement. Use "PARTIAL" for mixed results. Use "LOSS" for negative results. All three are valid findings.
Update DISCOVERIES.md if the finding answers an open question or establishes a new fact.
Not every experiment needs a changelog entry. After a PR is merged, the reviewing agent decides:
The changelog entry goes in docs/changelog.md with the next version number, a short description of the finding, and a link to the findings doc. The reviewing agent writes this on merge, not the contributor.
If the result is significant enough for a meeting presentation, use the prepare-meeting skill to compile it into a report.
