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 inclusionAI/Avernet --skill undercover-game-refereegit clone --depth 1 https://github.com/inclusionAI/AvernetWrote 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/inclusionai/avernet/undercover-game-referee)<a href="https://agentmods.dev/skills/inclusionai/avernet/undercover-game-referee"><img src="https://agentmods.dev/badge/skills/inclusionai/avernet/undercover-game-referee/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/inclusionai/avernet/undercover-game-referee"><img src="https://agentmods.dev/badge/skills/inclusionai/avernet/undercover-game-referee.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 | $0.00042 | $0.02141 |
| Opus 5 | $0.00021 | $0.01071 |
| Sonnet 5 | $0.00008 | $0.00428 |
| Haiku 4.5 | $0.00004 | $0.00214 |
Grade A, and why
undercover-game-referee 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 — 100 lines — stays where its author put it; the contents beside it link to each section on GitHub.
谁是卧底 · 当前动作
uc 表示 python3 skills/undercover-game-referee/scripts/undercover.py。
$session_id 从本次 GroupContext 获取，每局独立，命令显式带 --session "$session_id"。
正常执行只需本入口和当前节点正文；详细阶段/异常查 references/phase-machine.md，
参数查 references/commands.md，话术查 references/boards.md，拓扑与故障背景查 references/runs.md。
不预先通读这些参考文件。
BCS 身份与运行环境（本地、线上共用）
本次 BCS 投递的 GroupContext.recipient 是自己的正式 Bot ID，记为 $referee_uuid；
只有文本上下文时，取 BCS「你的身份」中的明确 ID。名称、发送者 ID、群发起人都不能替代。
缺少本次接收者 ID 时报告上下文缺失并结束激活，请维护者恢复投递信息。
begin 必须带 --referee-uuid "$referee_uuid"；脚本用 bcs-cli session get 获取当前会话，
核对自己是唯一 manager，玩家只取 worker。init 原样执行 begin 返回的命令，发牌前重新
核对名单并检查协作权限。--referee-uuid 仅用于 begin/init，不传给 open-*。
已有游戏的 status 返回 referee_uuid 时须与本次接收者 ID 一致；不一致立即停止。
所有 BCS CLI 操作直接使用运行环境提供的 bcs-cli；地址与认证由 CLI 和运行时管理。
保留已有环境，不设置本地地址、数据目录或 shell 别名。派遗言使用当前投递提供的
bcs_assign_task 调用方式；缺少该能力时报告环境未就绪并结束激活。
BCS_AUTH_FAILED / BCS_FORBIDDEN / 身份或名单不符：报告故障并结束激活，由维护者恢复 原 Bot 的认证、容器绑定或角色。认证失败不能靠新建会话解决。 主持人不执行 connect、不寻找或传递 token、不读源码或状态文件排障，也不改用底层命令绕过失败。 参数 BAD_ARGS 只按 commands.md 修正当前参数一次；其他失败按脚本明确给出的恢复动作处理， 没有恢复动作就停止。成功查询成员不代表认证成功，成功提交不代表玩家已经收到词或完成任务。
节点唤醒优先
收到 [State Machine Task] 就是 NODE_TASK，直到本次激活结束都不是 ECHO。
| node_id | 本次动作 |
|---|---|
| collect | 本阶段发言原话交节点指定的 speeches-set 命令（保留 --round/--stage/--attempt）；按返回 label/text 念遮蔽后的汇总，结束；不在这里 open-vote |
| tally | 全体存活玩家原始票面交节点指定的 votes-set 命令（保留 --round/--stage/--attempt）；结构化 human 票面保留为文本。pk 报票向与 PK 名单后结束，回灌再开 PK；continue 念开票稿后结束；finished 按下述终局步骤收尾 |
脚本校验当前阶段与运行尝试；沿用节点提供的 --round/--stage/--attempt，不能改参数套用迟到产物。节点命令报阶段不符时停止，不把迟到任务当作新阶段执行。
tally 刚判胜
uc reveal --session "$session_id"，仅本次 votes-set 返回 finished 时执行。- 主持稿以 reveal 返回的
finale_header原样开头，再公布词对、全员身份与词、胜负及关键转折；这是终局保密例外。 uc finish --session "$session_id"结束本会话，然后结束激活。失败如实报告。
该节点是 state_machine，不能使用 bcs_task_complete。禁止 bcs_route、路由给自己或寻找工具。 终局收尾就在当前 tally 内，不等回灌。只要不是本次刚判胜，FINISHED 不授权重复 reveal 或完成会话。
关闭失败恢复：维护者确认故障已修复并明确要求重试关闭时，核对本次 GroupContext 的会话 ID，仅执行 uc finish --session "$session_id"。此为 FINISHED 的恢复例外；不重复 reveal 或终局稿。只有 finish 成功才确认 BCS 会话已关闭。
What ships with it
16 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.
- references/boards.md 12 KB
- references/commands.md 18 KB
- references/phase-machine.md 19 KB
- references/runs.md 5.9 KB
- references/word-bank/easy.tsv 423 B
- references/word-bank/hard.tsv 457 B
- references/word-bank/medium.tsv 508 B
- scripts/undercover.py 112 KB runs code
- tests/fixtures/flow_baseline.json 43 KB
- tests/test_bcs_identity.py 6.4 KB runs code
- tests/test_finish.py 4.7 KB runs code
- tests/test_prompt_contracts.py 16 KB runs code
- tests/test_public_result.py 6.3 KB runs code
- tests/test_tie_break_pk.py 10 KB runs code
- tests/test_undercover_panel.py 8.0 KB runs code
- tests/test_vote_actor_ids.py 4.7 KB runs code
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 · +2 lines 95760bb69cc6
- 2d ago Changed · +24 lines 1020b29f60d5
- 3d ago First seen · 74 lines · 42 tokens per session scan A 5e73aff4c6d9
undercover-game-referee is a skill published in the GitHub repository inclusionAI/Avernet (547 stars, last pushed today), licensed Apache-2.0. It adds 42 tokens to every session and 2,141 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 2026-09-08.
Other skills, from other repositories
browser-a2d9b0b0
A browser gamepad skill for detecting connected controllers and reading their buttons and analogue sticks.
dsh-web-pet-developer
Create a pet for the dsh-pet plugin and integrate it into the dsh web GUI — author a v2 pet.json manifest plus an 8-column x 9-row atlas per the Codex/hatch-pet contract (live2d pets, voice packs and status decorations included), drop it into the pet-center user directory or contribute it as a built-in asset under…
reversi-master
One-click install + model switch:Claude Code,Codex CLI (OpenAI), Grok Build (xAI), DeepSeek Harness, Kimi Code (Moonshot) ,Qwen Code,Aider,OpenCode,MiMo Code (Xiaomi),ZCode (Z.AI),OpenClaw,Pi,OpenScience,Vibe-Trading,Claude Desktop (3P profile),ChatGPT desktop,OpenCode Desktop.
Trash Talk
AI becomes provocative and mocks the opponent during gameplay.
designer-perspective
A planning and decision-making skill that evaluates whether an action is justified, why it should be taken, and whether the available conditions are sufficient. It observes and recommends but does not perform operations or verify facts.
avalon-8-10p
Eight-to-ten-player Avalon rules for a permanent Agent host, with the user choosing to join the randomized player pool or spectate.