Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx agentmods add skills/ccdawn/codex-skills/briefbound-score-loopnpx skills add CCDawn/codex-skills --skill briefbound-score-loopgit clone --depth 1 https://github.com/CCDawn/codex-skillsWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/ccdawn/codex-skills/briefbound-score-loop)<a href="https://agentmods.dev/skills/ccdawn/codex-skills/briefbound-score-loop"><img src="https://agentmods.dev/badge/skills/ccdawn/codex-skills/briefbound-score-loop.svg" alt="Measured on agentmods" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.00049 | $0.01545 |
| Opus 5 | $0.00024 | $0.00772 |
| Sonnet 5 | $0.00010 | $0.00309 |
| Haiku 4.5 | $0.00005 | $0.00154 |
Grade A, and why
briefbound-score-loop scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 4d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 96 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Briefbound Score Loop
目标
围绕一个明确指标主动寻找候选,用可比较证据决定是否替换当前最好方案。它负责一条量化优化 lane,不负责整个研究方向或竞赛生命周期。
Briefbound task contract
- Context Boundary: 主指标、硬约束、当前最好方案、搜索空间、评价协议、预算和允许写入面。
- Output Contract: 下一候选、可比结果、决定、可复用证据和停止/转向信号。
- Allowed Action: 在已确认范围内提出并评估候选;不静默改变指标、数据、基线、预算或提交目标。
- Success Evidence: 精确 baseline/candidate、参数或 diff、命令、数据/seed、主副指标、运行产物和可比性判断。
- Stop Condition: 协议漂移、指标无法解析、候选重复、预算耗尽、连续结果不再提供新信息、写入冲突或安全边界变化。
- Route Out: 继续当前 loop、返回
briefbound-ai-research-loop、返回briefbound-competition-research-lifecycle、briefbound-bug-review、briefbound-router或 BLOCKED。
统一调用契约
- 只处理 Briefbound task contract;复合任务不吞其他 owner;Route Out 仅以 Briefbound task contract 为准。
- 用户可见内容默认中文,先说结论,再给比较依据。默认使用“替换、淘汰、继续观察、暂时无法比较”等通俗说法;必须保留枚举时写成中文结论加括号,例如“替换当前最好方案(
PROMOTE)”。 - 复杂指标、代理测试或搜索策略会影响判断时,用一句话解释;不展示内部 trial 账本。末行写
下一步建议: <一个具体动作>。
实验 owner 独占
只有主要未知量是“怎样持续改善一个明确指标”时进入。AI Research 中一次低成本比较由原 owner 直接执行。
分数下降、candidate reject 和 online neutral/worse 是实验结果,不是 TDD RED。metric/parser/schema/seed/shape/NaN/打包出现确定性错误时,临时路由 briefbound-bug-review,修复后回到原 lane。
Protocol Freeze
比较前冻结以下协议;任何一项变化都先建立新 baseline,不能把前后结果直接相减:
- 主指标、越大/越小更好和最小有效提升;
- 硬约束与副指标,不能用主指标掩盖合法性、正确性或资源超限;
- baseline 的 commit/hash/config;
- 数据、case、seed、运行环境和评价命令;
- 单候选预算、总预算和停止条件。
若目标确实有多个指标,先指定主指标与硬约束;只有用户明确接受综合规则时才做加权,不临时拼一个总分。
Search Policy
每轮先选择一种搜索意图,再生成候选:
EXPLOIT:已有稳定正向信号时,在当前最好方案附近做小步改动。EXPLORE:结果停滞或局部空间已重复时,尝试机制不同的方向。DIAGNOSE:噪声、代理指标或失败原因不清时,先设计能区分原因的检查。
把变量写成可判断的搜索空间:类别、顺序、数值或结构变化。候选必须有一个主要因果机制和唯一指纹;已做过的同机制、同方向、同边界候选直接跳过。随机变化若不能回答问题,也不进入队列。
ASK -> FILTER -> TELL
ASK:根据当前证据提出一个候选,写清机制、预期信号、smallestDecisiveEvaluation和 kill condition。FILTER:先做构建、合法性、成本和最小决定性检查。即使最乐观也过不了替换标准时,立即PRUNE;诊断候选则按是否回答问题判断。- 通过初筛后跑代表性评价;昂贵完整评价只留给仍可能晋升的候选。
TELL:记录结果、可比性、机制判断和副作用,再更新下一轮的EXPLOIT / EXPLORE / DIAGNOSE选择。- 没有新变量、新证据或校准价值时停止,不用更多试验掩盖平台期。
内部 trial 可标记 ASKED / RUNNING / PRUNED / COMPLETE / FAILED;用户通常只需要知道结果和原因。
决定规则
PROMOTE:可比较,主指标超过有效提升线,硬约束通过,代表性结果支持替换 baseline。REJECT:可比较但未过线,或副作用越过已确认边界。HOLD:方向有信息价值,但样本、稳定性或代表性不足,下一项验证明确。BLOCKED:协议漂移、证据缺失或结果不可解析,当前无法安全比较。
What ships with it
2 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 4d ago First seen · 96 lines · 49 tokens per session scan A 066c366a4fa0
briefbound-score-loop is a skill published in the GitHub repository CCDawn/codex-skills (4 stars, last pushed 22d ago), licensed MIT. It adds 49 tokens to every session and 1,545 once invoked, about $0.0002 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other skills, from other repositories
chinese-documentation
中文文档排版参考——中英文空格、全半角标点、术语保留、链接格式、中文文案排版指北约定。仅在用户显式 /chinese-documentation 时调用,不要根据上下文自动触发。.
chinese-git-workflow
国内 Git 平台配置参考——Gitee、Coding.net、极狐 GitLab、CNB 的 SSH/HTTPS/凭据/CI 接入差异与镜像同步配置。仅在用户显式 /chinese-git-workflow 时调用,不要根据上下文自动触发。.
chinese-code-review
中文 review 沟通参考——话术模板、分级标注(必须修复/建议修改/仅供参考)、国内团队常见反模式应对。仅在用户显式 /chinese-code-review 时调用,不要根据上下文自动触发。.
chinese-commit-conventions
中文 commit 与 changelog 配置参考——Conventional Commits 中文适配、commitlint/husky/commitizen 中文模板、conventional-changelog 中文配置。仅在用户显式 /chinese-commit-conventions 时调用,不要根据上下文自动触发。.
mcp-builder
MCP 服务器构建方法论 — 系统化构建生产级 MCP 工具,让 AI 助手连接外部能力.
systematic-debugging
Skill "systematic-debugging" from jnMetaCode/superpowers-zh, covering 系统化调试, 概述, 铁律, 何时使用 and 四个阶段.