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 masa-san-jp/Agent-Aiko --skill tailnet-exposegit clone --depth 1 https://github.com/masa-san-jp/Agent-AikoWrote 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/masa-san-jp/agent-aiko/tailnet-expose)<a href="https://agentmods.dev/skills/masa-san-jp/agent-aiko/tailnet-expose"><img src="https://agentmods.dev/badge/skills/masa-san-jp/agent-aiko/tailnet-expose/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/masa-san-jp/agent-aiko/tailnet-expose"><img src="https://agentmods.dev/badge/skills/masa-san-jp/agent-aiko/tailnet-expose.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 | $0.00172 | $0.01130 |
| Opus 5 | $0.00086 | $0.00565 |
| Sonnet 5 | $0.00034 | $0.00226 |
| Haiku 4.5 | $0.00017 | $0.00113 |
Grade C, and why
tailnet-expose scanned grade C with 2 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 10d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Asks for rootmediumPrivilege escalation
A mod that escalates privileges can change anything on the machine, not only the project.
## 仕組み(sudo 不要) Reads agent configuration directoriesmediumAgent snooping
.claude/, .codex/, .gemini/ hold keys, settings and other credentials a mod has no legitimate need for.
ネットワーク配信を伴うため、本スクリプトをエージェントの権限 allowlist に入れる。設定ファイル(例: `{{ORG_REPO_PATH}}/.claude/settings.json` の `permissions.allow`)に、自分の配置に合わせた絶対パスで登録する: What it actually says
/tailnet-expose
ローカルのもの(静的成果物・稼働中サービスのポート)を、tailnet 内だけの URL で出す汎用スキル。
何のためか
ユーザーが別端末(スマホ等)から成果物を見たい場合、file:// や localhost は別端末からは届かない。Tailscale の tailnet に出せば、同じ tailnet に属する端末からだけ URL で見れて、外部には出ない。これを「静的成果物」でも「稼働中サービス」でも 1 コマンドで出せる共通の土台にする。
2 つのモード
同梱スクリプト {{ORG_REPO_PATH}}/Agent-team/tools/tailnet-expose/tailnet-expose.sh を呼ぶ。
SK={{ORG_REPO_PATH}}/Agent-team/tools/tailnet-expose/tailnet-expose.sh
# 1) 静的ファイル/ディレクトリを配信
bash "$SK" serve <path-to-file-or-dir> [--port N]
# → http://<host>.<tailnet>.ts.net:<port>/[file]
# 2) 既にローカルで動いてるサービス(127.0.0.1:M)を tailnet に転送
bash "$SK" port <N> [--target M] # 既定 M=N
# → http://<host>.<tailnet>.ts.net:<N>/ (例: dev server localhost:3000 を port 3000 で出す)
bash "$SK" status # 出してるもの一覧
bash "$SK" stop [--port N|--all] # 片付け(使い終わったら必ず)
返ってきた URL をそのままユーザーの端末へ(チャット等で)送る。
仕組み(sudo 不要)
serve:python3 -m http.serverを tailscale IP に bind して配信する。port: 同梱のtcp_relay.py(依存なしの小さな TCP リレー)で tailscale IP:N → 127.0.0.1:M を中継する。- どちらも
tailscale serve/証明書/operator 設定/sudo は不要。bind 先を tailscale IP に限定するので LAN・インターネットには出ない。
不可侵ルール
- tailnet 内のみ。
tailscale funnel(インターネット公開)は絶対に使わない。 成果物には未公開情報が含まれうる。 - bind 先は tailscale IP に限定(
0.0.0.0/LAN IP に bind しない)。 - 使い終わったら
stop。 出しっぱなしにしない。 - 第三者の実名・機微は成果物自体に入れない。
有効化(allowlist のみ)
ネットワーク配信を伴うため、本スクリプトをエージェントの権限 allowlist に入れる。設定ファイル(例: {{ORG_REPO_PATH}}/.claude/settings.json の permissions.allow)に、自分の配置に合わせた絶対パスで登録する:
"Bash(bash {{ORG_REPO_PATH}}/Agent-team/tools/tailnet-expose/tailnet-expose.sh:*)"
検証状況
serve(静的配信)・port(既存サービス転送)とも tailscale IP 経由で HTTP 200 到達を実機検証済み。
関連
- HTML 特化版:
{{ORG_REPO_PATH}}/Agent-team/reference/skills/live-preview/SKILL.md(/live-preview。内部で本スキルのserveを呼ぶ) - 実装本体:
{{ORG_REPO_PATH}}/Agent-team/tools/tailnet-expose/tailnet-expose.sh
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.
- 10d ago First seen · 64 lines · 172 tokens per session scan C 6f19885d566e
tailnet-expose is a skill published in the GitHub repository masa-san-jp/Agent-Aiko (3 stars, last pushed 1mo ago), licensed MIT. It adds 172 tokens to every session and 1,130 once invoked, about $0.0009 per session on Opus 5. A static security scan graded it C with 2 findings (asks for root, reads agent configuration directories). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other skills, from other repositories
qwenpaw-usage
A QwenPaw command-line guide for setting up scheduled self-checks and packaging a QwenPaw workspace for another server.
aliyun-ecs
A command-line tool for managing Alibaba Cloud Elastic Compute Service (ECS), which provides rented virtual servers. It handles servers and related resources such as disks, snapshots, images, security groups, keys, and network interfaces.
storage-resize
A Linux procedure for expanding a cloud disk and then enlarging its partition and file system. It is designed for RPM-based systems such as Alibaba Cloud Linux and supports XFS, EXT4, and Btrfs.
backup-restore
A backup and recovery guide for projects, personal files, operating systems, disks, and Alibaba Cloud snapshots on command-line Linux systems that use the yum package manager. It first identifies the environment and then selects the relevant backup level.
floom
Create, run, and manage Floom AI workers — cloud automations that run on a schedule or a trigger and report back. Reach for this whenever the user wants something to happen repeatedly or in the background rather than a one-off script: a recurring task, "every day / every hour / every week", "schedule this", "run it on…
install-qwenpaw
An installation guide for running the QwenPaw AI assistant on Alibaba Cloud Linux 4, including setup for DingTalk and Alibaba Cloud's model service.