メニュー
Commit the current change, push a branch, and open a draft PR. Use when the user wants to open a PR for finished work.
Codex または Claude でインストール この Prompt をコピーして Codex、Claude、または他のアシスタントに貼り付けると、Skill ページを確認してインストールできます。
SOC 職業分類に基づく
Take a finished change from working diff to a ready-for-review PR, ending with the PR URL as plain text. Use when the user says "ship it" or wants to go from a working change to an opened PR.
How to drive the `ct` CLI to run matches and produce every number and graph a control experiment ships. Covers `ct run eval`, `ct run monitor`, `ct run summarize`, `ct game-metrics metrics / graph roc / graph pareto`, `ct traj`, and the viewers. Load this when you launch a match, compute safety or usefulness, regenerate a graph, or need the exact reproduce command for an output. White loads it always; main and others load it on demand.
The empirical-research method and quality bar for AI-control experiments run through the ct CLI. Reduce uncertainty fast, iterate on qualitative plus quantitative evidence, size for power, keep splits clean, report honestly, and hold P1 observability, P2 affordance discipline, and the per-team B/R/W/M criteria. Loaded by every role.
The public, LessWrong-style writeup of a control experiment. Ruthlessly compressed for a reader with zero context, leading with the surprise, citing public-viewer trajectory URLs only. Load this in the writing phase when producing the shipped post into `outputs/`. It lifts the STRONG-tagged claims from the internal [report-writing]; same evidence base, far less detail. Before shipping it runs the lw-review gate below.
The red (attacker) role for ONE protocol in an automated control-research run. Load this when you are spawned as red_<protocol>. Holds red's self-managed loop, guarded objective, iteration discipline, and validity rules. Every shared fact (folder layout, iteration mechanics, the stop gate, the game, safety formula, the quality bar, ct commands) lives in a general skill and is referenced here by name, never copied.
The white (referee) role for ONE protocol in an automated control-research run. Load this when you are spawned as white_<protocol>. Holds white's self-managed loop, guarded objective, task-set ownership, validity disciplines, and per-protocol stopping judgment. Every shared fact (folder layout, iteration mechanics, the stop gate, the game, safety formula, the quality bar, ct commands) lives in a general skill and is referenced here by name, never copied.
| name | pr |
| description | Commit the current change, push a branch, and open a draft PR. Use when the user wants to open a PR for finished work. |
make check green locally first (skip the wait if no Python changed). If the change alters commands, workflows, or behavior, update the relevant docs/ guide in the same diff.main, but case-dependent (e.g. stack onto an open PR's branch when the work builds on it). Commit your files only, message formatted as the title below plus the Co-Authored-By trailer. History stays linear: rebase, never merge commits.gh pr create --draft with the title, description, and label below.Conventional Commits: type(scope): description. We squash-merge: the PR title becomes the commit subject on main and feeds the release notes.
pr-title-lint): feat, fix, docs, refactor, perf, test, chore, ci, build, revert, or research (research PRs use scope blue or red).cli, eval, monitor, protocols, trajectories, sandbox, fleet, environments, viewer, live. Omit for cross-cutting work.! after type or scope for a breaking change.What and why, kept accurate as the branch changes. Link relevant runs (https://equistamp.linuxarena.ai/runs/<run_id>) and trajectories (https://equistamp.linuxarena.ai/trajectories/<trajectory_id>).
Add the matching label — it groups the release notes: tooling, environment, straj, main-task / side-task, bug, documentation.
Don't force-push, merge, close, or post public comments without the user asking.