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 skills add hxy91819/mason-skills --skill bb-stewardgit clone --depth 1 https://github.com/hxy91819/mason-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/hxy91819/mason-skills/bb-steward)<a href="https://agentmods.dev/skills/hxy91819/mason-skills/bb-steward"><img src="https://agentmods.dev/badge/skills/hxy91819/mason-skills/bb-steward/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/hxy91819/mason-skills/bb-steward"><img src="https://agentmods.dev/badge/skills/hxy91819/mason-skills/bb-steward.svg" alt="Reviewed on agentmods" width="80" 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 5.1 | $0.00055 | $0.02206 |
| Opus 5.5 | $0.00022 | $0.00882 |
| Sonnet 5.5 | $0.00011 | $0.00441 |
| Haiku 4.5 | $0.00006 | $0.00221 |
Grade A, and why
bb-steward 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 today.
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 — 86 lines — stays where its author put it; the contents beside it link to each section on GitHub.
BB 管家
这是流程类 Skill,仅在用户显式调用 $bb-steward 时进入。管家是会话里的编排者:自己决定下一步、把具体工作派给
BB 子线程、收通知后汇总,并把跨会话必须保留的信息写进工作簿。
分工是判断留下,调研和体力活派出。要懂用户意图和来龙去脉才能做对的判断(范围、归属、取舍、做到什么程度)由管家 自己做。判断要用的事实,几条命令能查完的管家自己查;要大量阅读或排查的,例如读代码和文档、上服务器翻日志找 bug 线索、收集舆情和外部信息,派只读调研线程:只给问题、不附管家的猜测,要它带证据交回事实和可能原因,不替管家定怎么做, 不改东西。代码和文档改动、长时间运行、可并行的独立任务、独立审查和经手秘密的工作也 派出去。每转一手都会丢细节,子线程摘要里的错误又和正确结论写得一样笃定,所以判断不外包,验收不只看摘要。
开始前加载 $bb-model-routing,派发、难度归类、任务模板和等待规则全部以它为准,本 Skill 不重复。
与 large-task 的分工
管家处理"跨会话、但不值得拆 Story"的目标:人在环路,Agent 自己拿主意,没有脚本状态机。目标大到需要 Story
拆分、无人值守连续执行时,转 $large-task-planning 建计划,再由 $large-task-orchestrator 执行;工作簿的路线
进 SPEC 的交付路线,关键决策进 SPEC 的已定取舍,用户待办进 STATUS 的需要关注,然后删除工作簿。两套不并存。
工作簿
每个目标一份,提交进目标仓库,默认路径 docs/steward/<topic>.md;项目 AGENTS.md 另有指定时以项目为准。
工作簿、项目 AGENTS.md 里指向它的那一行("存在时先读")以及项目文档索引的登记,都算工作簿本身,由管家
直接写并按项目约定提交,不派子线程;它们让不经本 Skill 进入的会话也能接上。其余改动仍派出去。子线程只在
"先读"里读工作簿。格式见工作簿模板。
工作簿只有三段,各自的准入判据:
- 路线:阶段清单和当前位置,一个阶段一行,标出已完成、进行中、未开始。项目已有路线图(issue、计划文档) 时只放链接和当前位置,不复制阶段。任务清单、"做了什么"、子任务矩阵、实时配额计数和中间 commit SHA 不进这里,那是项目 Issue、派发文本和 git log 的事。
- 关键决策:两条同时成立才记:下个会话不知道这条就很可能重做或翻案;从代码、git、既有文档推不出来。一条 一行,写决定、日期、原因、影响范围。新决定覆盖或收窄旧决定时,直接原地更新该行,不按时间流水账追加互相冲突的 旧条目。项目已有决策登记制度(决策地图、ADR)时只放链接,这段剩下的通常只有管家与用户之间的操作约定,例如 哪些事用户亲自裁决、合并前固定派什么审查。子线程回复里"自行引入的决定"逐条过这个判据,绝大多数随线程消亡。
- 用户待办:只有用户本人能做的事,每条写做什么、为什么只能用户做、挡住了哪个阶段、提出日期、状态。 不挡任何阶段的事"挡住"栏写"不挡",排在挡阶段的后面。待办只有待办和已完成两种状态;用户说做完或管家 验证到即改为已完成,保留原行。
工作簿是人读的状态页,不是实时运行日志。只在阶段完成、路线调整、关键决策变更或用户待办状态变更时更新并提交工作簿; 子线程启动、中途进度、配额消耗计数不改工作簿。每次修改工作簿后运行确定性门禁校验结构与一屏容量(三段标题固定、非空行 $\le 45$、单行可见字符 $\le 180$、路线禁塞表格):
python3 scripts/steward_workbook.py --mode check --file <工作簿路径>
python3 scripts/steward_workbook.py --mode should-commit --file <工作簿路径>
一段超过一屏或 --mode check 报错就说明准入判据没守住,先删再加;--mode should-commit 返回 should_commit: false 时说明仅有括号内微状态描述调整,不单独生成 git commit。
会话开始
- 读工作簿;读
bb status,检查分支、git status --short、git worktree list。没有工作簿时先建:从bb status的子线程、当前会话上下文、git log 和项目既有计划恢复路线位置和已埋的用户待办,再继续。 - 从路线的当前位置决定下一步;遇到用户依赖时按下节处理。
What ships with it
4 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.
- today Changed · +16 lines 11d8ab5383f6
- 2d ago Changed · -2 lines e6979cfea8e6
- 6d ago Changed 41db15a064f0
- 10d ago Changed · -5 lines 0eac03b679d2
- 11d ago Changed · +10 lines 60fd29d7cc0b
- 13d ago First seen · 67 lines · 55 tokens per session scan A 9e1f1ac14b4b
bb-steward is a skill published in the GitHub repository hxy91819/mason-skills (3 stars, last pushed yesterday), licensed MIT. It adds 55 tokens to every session and 2,206 once invoked, about $0.0002 per session on Opus 5.5. 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 2026-10-13.
Other skills, from other repositories
story-readiness
Checks whether a story is ready to implement: clear acceptance criteria, no open questions, valid ADR and dependencies. Use when the user asks if a story, or the sprint's stories, are ready to start or build.
recipe-create-meet-space
Create a Google Meet meeting space and share the join link.
atmos-config
Atmos root configuration: atmos.yaml discovery, precedence, deep merging, basepath, imports, minimal bootstrap, and routing to narrower Atmos skills.
workthreads
SpecStory Workthreads - a weekly work-thread rollup across a team's repos from SpecStory coding histories (any agent - Claude Code, Codex, Cursor, Gemini, and more). It groups the window's sessions into threads of work per project and labels each new / open / recently closed, so a lead sees what shipped, what is still…
qa-project-bootstrap
Onboard a new QA engineer to an existing codebase, or audit an existing test architecture. Produces a 30-day ramp plan: codebase orientation, framework walkthrough, test architecture audit, mentorship pairing, and first-test guidance. Use when: "QA onboarding," "new tester," "ramp up," "test architecture audit,"…
remove
Remove a deployed framework or addon from the current workspace.