菜单
Handle the two situations the greedy ratchet cannot: route safety-critical or high-risk changes to human review, and switch search strategy when the loop plateaus (greedy local search -> grid / tree / evolutionary). Use from inside a ratchet-experiment loop.
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
Run a disciplined keep/revert experiment loop against a frozen scalar evaluator. Use when iterating on a model, algorithm, or piece of code to find improvements: propose a change, run a bounded experiment, score it, and keep the change only if it improves the metric (otherwise revert). Requires a frozen evaluator that emits a single scalar plus hard-constraint vetoes.
For an accepted improvement, attribute the gain to the specific component via ablation, and confirm it is real (not noise) with multi-seed re-runs, before trusting or shipping it. Use after ratchet-experiment keeps a change.
Read a paper or product writeup (and its key references) and distill it into a structured, implementable spec: the core mechanism, the claimed metric, a minimal reproduction, and the constraints. Use at the start of adopting a community idea, before reproducing or grafting it.
Integrate a confirmed idea into an existing large codebase behind tests — the step generic paper-to-code tools skip. Use after repro-harness, before validating with ratchet-experiment.
Confirm a distilled idea actually works in isolation, against a frozen evaluator, before grafting it into a real codebase. Guards against fabricated or hardcoded results. Use after distill-paper and before graft-change.
| name | escalate |
| description | Handle the two situations the greedy ratchet cannot: route safety-critical or high-risk changes to human review, and switch search strategy when the loop plateaus (greedy local search -> grid / tree / evolutionary). Use from inside a ratchet-experiment loop. |
Status: v0 scaffold — unverified. Behavior not yet validated end-to-end in this repo. See
../../STATUS.md.
The keep/revert ratchet is greedy and autonomous. Two conditions require escalation.
If a proposed change touches a safety-relevant path (planner, controller, anything that could cause real-world harm) or is otherwise high-risk:
../../AGENTS.md.If N consecutive experiments fail to improve the metric beyond measured noise, the
greedy ratchet is likely stuck in a local optimum (see
../../docs/pitfalls.md). Switch strategy:
../../src/talos/orchestration.py).Either a human-review request (Trigger 1) or a switch to a broader search procedure (Trigger 2), recorded in the ledger so the change of strategy is auditable.