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 b3rys/b3rys-team-os --skill b3os-model-rolloutgit clone --depth 1 https://github.com/b3rys/b3rys-team-osWrote 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/b3rys/b3rys-team-os/b3os-model-rollout)<a href="https://agentmods.dev/skills/b3rys/b3rys-team-os/b3os-model-rollout"><img src="https://agentmods.dev/badge/skills/b3rys/b3rys-team-os/b3os-model-rollout/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/b3rys/b3rys-team-os/b3os-model-rollout"><img src="https://agentmods.dev/badge/skills/b3rys/b3rys-team-os/b3os-model-rollout.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 | $0.00157 | $0.01396 |
| Opus 5.5 | $0.00063 | $0.00558 |
| Sonnet 5.5 | $0.00031 | $0.00279 |
| Haiku 4.5 | $0.00016 | $0.00140 |
Grade B, and why
b3os-model-rollout scanned grade B with 1 finding 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 14d 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.
Reads agent configuration directoriesmediumAgent snooping
.claude/, .codex/, .gemini/ hold keys, settings and other credentials a mod has no legitimate need for.
- 공유 설정(호스트 `~/.codex/config.toml`, openclaw `agents.defaults`)을 바꾸는 것. How it starts
The opening of the file, as written. The whole thing — 60 lines — stays where its author put it; the contents beside it link to each section on GitHub.
b3os-model-rollout — 팀원 모델 전환
새 모델이 나올 때마다 런타임마다 설정 파일·재시작 방법·확인 방법이 달라 손으로 하던 일을 한 도구로 묶었다. 결정(어느 모델로 바꿀지)은 사람이 한다. 이 도구는 "되는지 재고, 바꾸고, 확인하고, 안 되면 되돌린다" 만 한다.
명령
S=skills/b3os-model-rollout/scripts/model-rollout.ts
bun $S inventory # 팀원별 런타임 · 지금 모델 · effort · 설정 파일
bun $S check --model gpt-6.1-sol --effort medium # 실제로 되는지(설정 안 바꿈). 전원 works 면 exit 0
bun $S apply --model gpt-6.1-sol --effort medium # 미리보기 — 아무것도 바꾸지 않는다
bun $S apply --model gpt-6.1-sol --effort medium --yes # 실제 전환
# 공통: --members a,b(일부만) · --json · --registry <agents.json>
# 다른 맥: --host <ssh 이름> --remote-repo <그 맥의 저장소 경로> (그 맥 저장소의 같은 스크립트를 돌린다)
런타임별로 무엇을 보고 무엇을 바꾸나
| 런타임 | 팀원 목록 | 설정 | 확인(check) | 재시작 |
|---|---|---|---|---|
| hermes | agents.json 의 hermes_profile |
~/.hermes/profiles/<p>/config.yaml — model: 블록 default, agent: 블록 reasoning_effort |
one-shot 1회(--provider 필수) → usage 파일의 실제 model |
그 팀원 게이트웨이(gateway_service 또는 ai.hermes.gateway-<p>) |
| codex | agents.json 의 runtime: codex |
~/.codex-agents/<id>/config.toml — 최상위 model·model_reasoning_effort |
codex exec 1회 → 세션 기록의 실제 model |
그 팀원 브리지(<prefix>.codex-bridge-<id>) |
| openclaw | ~/.openclaw/openclaw.json 의 agents.entries(모델을 따로 정한 에이전트) |
openclaw config set agents.entries.<id>.model |
게이트웨이 openclaw models list 의 available === true |
게이트웨이 1회(openclaw 팀원 전원 잠깐 멈춤) |
- claude 런타임은 대상이 아니다 — 모델은 CLI 쪽 설정이다.
- codex 실행 파일:
CODEX_BIN이 없으면 저장소.env의CODEX_BIN한 줄을 쓴다. 새로 받은 Homebrew 사본은 macOS 격리 표시 때문에 화면 없이 실행하면 멈출 수 있다 — 그럴 때는 격리 표시가 없는 사본을CODEX_BIN으로 준다.
판정 규칙
works= 실제 호출이 그 모델로 응답했다(openclaw 는 게이트웨이가 available=true).not_supported= 다른 모델이 응답했거나, 게이트웨이가 그 모델을 모른다.unknown= 인증·한도·시간 초과 — "지원 안 됨" 이 아니다. 원인부터 푼다.- 목록(카탈로그)에 이름이 있다는 것은 근거가 아니다. 구독 로그인에서 실제로 되는지는 check 로만 안다.
apply 가 하는 일
- 이미 그 모델이면 건너뛴다.
- 설정 모양이 예상과 다르면 그 팀원은 건너뛰고 "손으로" 라고 적는다(줄을 새로 끼워 넣지 않는다).
- 백업(
<설정>.bak-rollout-<시각>) → 값 한 줄 → 그 팀원만 재시작 → 설정만으로(덮어쓰기 없이) 다시 check. - 확인이 실패하면 백업 복원 + 재시작.
- openclaw: 먼저 게이트웨이가 모델을 아는지 본다. 모르면 아무도 바꾸지 않는다. 알면
openclaw.json백업 → 에이전트마다config set→ 게이트웨이 재시작 1회 → 반영 확인, 실패하면 복원 + 재시작.
What ships with it
1 file 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.
- 14d ago First seen · 60 lines · 157 tokens per session scan B f80826887dc3
b3os-model-rollout is a skill published in the GitHub repository b3rys/b3rys-team-os (14 stars, last pushed yesterday), licensed Apache-2.0. It adds 157 tokens to every session and 1,396 once invoked, about $0.0006 per session on Opus 5.5. A static security scan graded it B with 1 finding (reads agent configuration directories). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-10-02.
Other skills, from other repositories
fable-commander
A planning and review workflow in which one model prepares decisions while selected models handle research and execution. It is written for Fable and Claude Code, tools for coordinating multi-step software work.
takt
A workflow engine that runs a task through a sequence of steps defined in a YAML file, using separate coding agents for the work.
codebase-context-extractor
This skill provides a comprehensive context extraction system for large codebases. It intelligently analyzes code structure, dependencies, and relationships to extract relevant context for understanding, debugging, or modifying code.
skill-creator
Generates Anthropic Skills with complete workflow including GitHub PR creation and local download verification.
Research Synthesis Workflow
A step-by-step guide to synthesizing research from multiple sources into a coherent summary.
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.