跳到正文
FunCoding

搜索

搜索文档、智能体、博客、Skill 和 MCP

add-dial-tool

Give chosen NanoClaw agents a real phone number as a container tool — the `dial` CLI baked into the agent image plus OneCLI credential injection for api.getdial.ai, scoped per agent, so the agents you pick can send SMS, place AI voice calls, and receive verification codes from inside the sandbox. Independent of the Dial channel; idempotent; re-run to change which agents may use it. Use when the user wants agents to text, call, or run `dial …` from a chat, without wiring Dial as a messaging channel.

AI 与智能体31k.claude/skills/add-dial-tool/SKILL.md

安装

将以下指令发送给 Claude Code、Codex 或 Cursor,智能体会先检查内容的安全性,经你确认后再安装。

读取 https://funcoding.ai/skills/nanocoai/nanoclaw/add-dial-tool/install.md ,按里面的步骤帮我安装这个 Skill。

SKILL.md

Add Dial Tool

Installs Dial as a container tool: the dial CLI on the agent's PATH, the dial-cli skill so the agent knows how to drive it, and an OneCLI credential so in-container calls are injected keyless. Independent of the Dial channel (/add-dial) — install this alone. Idempotent: re-run it to change which agents may use Dial.

This tool spends money and reaches real people. An agent with Dial access can text and call any number and buy more numbers, billed to the Dial account. The CLI and the skill file land in every agent's container, but the key is injected per agent by OneCLI, so the operator chooses which agents get it. Every other agent gets an OneCLI block rule and sees 403 blocked_by_policy if it tries.

Run this from the NanoClaw repo on the host (not from a chat with an agent — the container can't install itself). The mechanical steps carry nc: directive fences: an agent reads the prose and applies them, and a parser can apply them deterministically from the same document. Every directive is idempotent, so the whole skill is safe to re-run; anything a parser can't apply falls back to the prose beside it.

Pre-flight

OneCLI is required for credential injection — without it there is no way to hand the key to a container without putting it in an env var. This must succeed before anything else runs:

command -v onecli >/dev/null

If it fails, tell the user to run /init-onecli first, then retry. Stop here.

This skill scopes Dial with legacy OneCLI block rules. OneCLI gateway 1.42 and later reject legacy rule writes (410), and that includes 1.42.0, the version NanoClaw pins. Writing the Dial key first and then failing on the rules would leave every all-mode agent able to use Dial. So read the version of the gateway the onecli CLI talks to before anything is written, and stop unless it is older than 1.42. The Dial sign-up and credential steps below use the captured version, so they cannot run when this check fails:

U=$(onecli config get api-host | jq -r '.value // empty') && [ -n "$U" ] || { echo "could not read the onecli CLI's api-host, so the OneCLI gateway version cannot be checked. Nothing was written to OneCLI." >&2; exit 1; }; H=$(curl -fsS --max-time 10 "$U/api/health") || { echo "could not reach the OneCLI gateway at $U. Nothing was written to OneCLI." >&2; exit 1; }; V=$(printf '%s' "$H" | jq -er '.version') || V=; case "$V" in 0.*|1.[0-9].*|1.[1-3][0-9].*|1.4[01].*) printf '%s\n' "$V" | grep -Eq '^[0-9]+\.[0-9]+\.[0-9]+$' && { echo "$V"; exit 0; };; esac; echo "OneCLI gateway version '${V:-unreadable}' at $U: /add-dial-tool needs legacy OneCLI rules, which gateway 1.42 and later reject. Nothing was written to OneCLI. This skill is not supported on the pinned gateway (1.42.0) until it moves to the OneCLI policy API." >&2; exit 1

If it fails, show the user the message as it is and stop. Do not work around it by turning off OneCLI policy enforcement or granting the Dial key by hand: the block rules are what keep agents you did not choose away from Dial.

Calls this setup makes to Dial identify the install. The dial CLI prepends DIAL_USER_AGENT to its own token, so the account's requests stay attributable to this NanoClaw install in Dial's server-side logs. Resolve the token once (nanoclaw/<version>; an unreadable package.json degrades to nanoclaw/unknown rather than blocking the install):

node -p "'nanoclaw/'+(require('./package.json').version||'unknown')" 2>/dev/null || echo nanoclaw/unknown

Prefix every dial command below with DIAL_USER_AGENT={{dial_ua}}.

Choose which agents may use Dial

List the agent groups (the NanoClaw service must be running — ncl talks to it over its socket):

ncl groups list --json | jq -r 'if (.data|length)==0 then "no agent groups yet" else [.data[] | "\(.id) (\(.name))"] | join(", ") end'

Ask the operator which of them may use Dial. Say plainly what they are granting, and ask even when there is a single agent:

Agents on this install: {{agent_groups}}. Giving an agent Dial lets it text and call any number and buy numbers, billed to your Dial account. Agents you leave out are blocked at the gateway (reversible by running /add-dial-tool again). Agents created after this run have Dial until the next run.
Which agents may use Dial? Enter agent ids separated by commas with no spaces (the `ag-…` column), `all` for every agent, or `none` to install the tool with every agent blocked for now.

all and none cannot be mixed with ids, and an empty answer is never "everyone". A typo must not silently open or close anything, so every id named must be a real agent group:

for w in $(printf '%s' '{{dial_agents}}' | tr ',' ' '); do case "$w" in all|none) ;; *) ncl groups list --json | jq -e --arg id "$w" '.data[] | select(.id==$id)' >/dev/null || { echo "unknown agent group '$w' — see: ncl groups list" >&2; exit 1; }; esac; done

Install the Dial CLI on the host

The host needs the dial CLI to sign in: dial auth login / dial auth verify-otp write the host auth file that the credential step below reads. Pinned to the same version the agent image gets, so host and sandbox agree:

command -v dial >/dev/null || npm install -g @getdial/cli@0.37.0

Sign in to Dial

Dial's CLI owns the account credential (an auth file it writes on sign-in).

Check the host sign-in

Is this host already signed in?

DIAL_USER_AGENT={{dial_ua}} dial doctor --json

Read the account

If it is, read which account — that account's key is what the chosen agents will use:

DIAL_USER_AGENT={{dial_ua}} dial doctor --json
This host is signed in to Dial as {{connected_email}}; the agents you chose will use that account. To give them a different account, run `dial auth login <email> --force` and `dial auth verify-otp --code <code>` on the host first, then run /add-dial-tool again.

Send the code

If it is not, verify an email with a one-time code. Collect the email:

What's your email? Dial sends a one-time code to verify it. By continuing you create a Dial account and agree to Dial's Terms of Service (https://getdial.ai/terms) and Privacy Policy (https://getdial.ai/privacy).

Send the code (--force re-sends even if a prior code is pending):

: "OneCLI gateway {{onecli_gateway}}"; DIAL_USER_AGENT={{dial_ua}} dial auth login {{owner_email}} --force

Verify the code

Collect the code:

Enter the 6-digit code from your email

Verify it. Do not pass --agent nanoclaw here: this skill owns the container dial-cli skill, and --agent would drop a second, unmanaged copy next to it:

: "OneCLI gateway {{onecli_gateway}}"; DIAL_USER_AGENT={{dial_ua}} dial auth verify-otp --code {{otp}}

Put the CLI and its skill in the agent image

The agent's global Node CLIs install from container/cli-tools.json, not from hand-edited Dockerfile layers. Add the pinned Dial CLI — idempotent on name, so a re-run is a no-op. @getdial/cli has no native postinstall, so no onlyBuilt:

{ "name": "@getdial/cli", "version": "0.37.0" }

The version (0.37.0) is the canonical pin — this document is the source of truth; the host install above uses the same one.

Mount the sandbox-aware dial-cli skill so the agent knows the CLI runs keyless in there and never asks for credentials. container/skills/ is mounted read-only into every agent container (at /app/skills) — which is why the key, not the skill file, is what gets scoped per agent:

container-skills/dial-cli/SKILL.md -> container/skills/dial-cli/SKILL.md

Rebuild the image so the CLI lands. On an install that fetches a published image this adds Dial as a layer on top of it; on one that builds its own it rebuilds:

./container/build.sh

Register the credential with OneCLI

Read the API key from the host auth file — the single source of truth, written by dial auth login / dial auth verify-otp — and put it in the OneCLI vault for api.getdial.ai. Always replace: the vault is keyed by name, so an existing "Dial API" secret is not necessarily this account's (re-onboarding, switching accounts, or rotating the key all leave a secret whose value points at the previous account, and a sandboxed agent then lists that account's numbers). A stale secret is deleted and a fresh one created rather than updated in place: onecli secrets update accepts a new value only on the command line, and the key must never sit on one. It travels through a 0600 temp file that is removed right after (--file), so it is never on argv or in a captured variable. Selective-mode agents pick the new id up in the merge step below:

: "OneCLI gateway {{onecli_gateway}}"; T=$(mktemp) && chmod 600 "$T" && jq -r '.apiKey // empty' "${XDG_DATA_HOME:-$HOME/.local/share}/dial/auth.v1.json" > "$T" 2>/dev/null; [ -s "$T" ] || { rm -f "$T"; echo "no Dial API key in the host auth file — sign in with dial auth login / verify-otp, then re-run" >&2; exit 1; }; S=$(onecli secrets list | jq -r 'first(.data[] | select(.name | test("(?i)dial"))) | .id // empty'); if [ -n "$S" ]; then onecli secrets delete --id "$S" >/dev/null || { rm -f "$T"; echo "could not remove the previous Dial secret $S" >&2; exit 1; }; fi; onecli secrets create --name "Dial API" --type generic --file "$T" --host-pattern api.getdial.ai --header-name Authorization --value-format "Bearer {value}" >/dev/null; rc=$?; rm -f "$T"; exit $rc

Scope it to the chosen agents

Create the OneCLI agents

NanoClaw gives every agent group its own OneCLI agent whose identifier is the group id, created on the group's first spawn. A group that has never spawned has no OneCLI agent yet, and a block rule needs one to attach to — so create the missing ones now, exactly as the runtime would (secret mode all, nothing else touched):

G=$(ncl groups list --json) || { echo "could not list agent groups — is the NanoClaw host running?" >&2; exit 1; }; AG=$(onecli agents list) || { echo "could not list OneCLI agents" >&2; exit 1; }; printf '%s' "$G" | jq -r '.data[] | "\(.id)\t\(.name)"' | while IFS="$(printf '\t')" read -r gid gname; do printf '%s' "$AG" | jq -e --arg g "$gid" '.data[] | select(.identifier==$g)' >/dev/null || onecli agents create --name "$gname" --identifier "$gid" >/dev/null || { echo "could not create an OneCLI agent for $gname ($gid)" >&2; exit 1; }; done

Set the block rules

The one switch is a per-agent block rule on api.getdial.ai, named Dial: blocked for <group> so only this skill's rules are ever read or written (an operator's own rules on the host are left alone). A chosen agent has its rule removed; every other agent has one present and enabled. A 403 blocked_by_policy in a container means "not chosen", not "broken":

A=$(printf '%s' '{{dial_agents}}' | tr -d ' '); G=$(ncl groups list --json) || { echo "could not list agent groups — is the NanoClaw host running?" >&2; exit 1; }; case ",$A," in *,all,*) A=$(printf '%s' "$G" | jq -r '[.data[].id] | join(",")');; esac; AG=$(onecli agents list) || { echo "could not list OneCLI agents" >&2; exit 1; }; RL=$(onecli rules list) || { echo "could not list OneCLI rules" >&2; exit 1; }; printf '%s' "$G" | jq -r '.data[] | "\(.id)\t\(.name)"' | while IFS="$(printf '\t')" read -r gid gname; do aid=$(printf '%s' "$AG" | jq -r --arg g "$gid" 'first(.data[] | select(.identifier==$g)) | .id // empty'); [ -n "$aid" ] || { echo "no OneCLI agent for $gname ($gid)" >&2; exit 1; }; rid=$(printf '%s' "$RL" | jq -r --arg a "$aid" 'first(.data[] | select(.hostPattern=="api.getdial.ai" and .action=="block" and .agentId==$a and (.name | startswith("Dial: blocked for ")) and ((.pathPattern // "")=="") and ((.method // "")==""))) | .id // empty'); case ",$A," in *,"$gid",*) if [ -n "$rid" ]; then onecli rules delete --id "$rid" >/dev/null || { echo "could not remove the Dial block for $gname ($gid)" >&2; exit 1; }; fi; echo "allowed: $gname ($gid)";; *) if [ -z "$rid" ]; then onecli rules create --name "Dial: blocked for $gname" --host-pattern api.getdial.ai --action block --agent-id "$aid" --enabled >/dev/null || { echo "could not create the Dial block for $gname ($gid)" >&2; exit 1; }; else onecli rules update --id "$rid" --enabled true >/dev/null || { echo "could not re-enable the Dial block for $gname ($gid)" >&2; exit 1; }; fi; echo "blocked: $gname ($gid)";; esac; done

Merge secrets for selective agents

Secret lists are left alone, with one exception. An agent in selective mode only gets the secrets on its list, so a chosen selective agent has the Dial secret merged into it. onecli agents set-secrets switches an agent to selective mode, so it is never called on an all-mode agent — that would silently cut the agent off from every credential not on its list. Blocked agents keep their lists untouched in either mode; the rule alone blocks:

A=$(printf '%s' '{{dial_agents}}' | tr -d ' '); case ",$A," in *,all,*) A=$(ncl groups list --json | jq -r '[.data[].id] | join(",")');; esac; S=$(onecli secrets list | jq -r 'first(.data[] | select(.name | test("(?i)dial"))) | .id // empty'); [ -n "$S" ] || { echo "no Dial secret in the OneCLI vault — the credential step above did not complete" >&2; exit 1; }; onecli agents list | jq -r '.data[] | select(.secretMode=="selective") | "\(.id)\t\(.identifier)"' | while IFS="$(printf '\t')" read -r aid gid; do case ",$A," in *,"$gid",*) onecli agents set-secrets --id "$aid" --secret-ids "$(onecli agents secrets --id "$aid" | jq -r --arg s "$S" '[.data[], $s] | unique | join(",")')" >/dev/null || { echo "could not add the Dial secret to $gid" >&2; exit 1; }; echo "Dial secret added to the list of $gid";; esac; done

Hand the tool to running agents

container/skills/ is mounted read-only into every agent container, and each group's .claude-shared/skills/ holds symlinks into that mount that are synced when the container spawns — so nothing is copied per session. A running agent keeps its old image until it respawns, so restart every group; without a --message each one comes back on its next message, on the new image, with the CLI on PATH and the skill in place. This is a restart effect, so it does not fire after an earlier step bounced — agents keep the image they have until the gap above is fixed and the skill is re-applied:

ncl groups list --json | jq -r '.data[].id' | while read -r gid; do ncl groups restart --id "$gid" >/dev/null || { echo "could not restart $gid" >&2; exit 1; }; done

Done

The chosen agents can now use Dial from inside their containers; the others are blocked at the gateway. Auth is injected by OneCLI; a 403 blocked_by_policy means the agent was not chosen (run /add-dial-tool again to change that); a 401 means the Dial secret needs (re)connecting — not a login. Verify from a chat with a chosen agent: "run dial doctor" or "text +1… hi".

To uninstall: see REMOVE.md. To wire Dial as a messaging channel too, run /add-dial.

Troubleshooting

command -v onecli fails. OneCLI is not installed or not on PATH. Run /init-onecli, then re-run this skill.

ncl can't reach the host. The agent list and the scoping steps talk to the running NanoClaw service. Start it (pnpm run dev, or restart the service) and re-run.

unknown agent group. An id in your answer is not in ncl groups list. Copy the ag-… id exactly; names are not accepted.

no Dial API key in the host auth file. The sign-in did not complete. Run dial auth login <email> --force, then dial auth verify-otp --code <code>, and re-run.

A chosen agent gets 401. The vault secret is stale (a different account's key, or a rotated one). Re-run this skill — it always rewrites the secret with the key the host is signed in with.

An agent you left out can still use Dial. It was created after the last run (a new OneCLI agent starts in all mode with no rule). Re-run this skill; it only touches the per-agent rules.

dial: command not found inside a container. The image predates the manifest entry. Run ./container/build.sh, then ncl groups restart --id <group-id> so the agent respawns on it.

相似的 Skill

mcp-builder
官方
anthropics/skills180k

mcp-builder

Guide for creating high-quality MCP (Model Context Protocol) servers that enable LLMs to interact with external services through well-designed tools. Use when building MCP servers to integrate external APIs or services, whether in Python (FastMCP) or Node/TypeScript (MCP SDK).

AI 与智能体

internal-comms
官方
anthropics/skills180k

internal-comms

A set of resources to help me write all kinds of internal communications, using the formats that my company likes to use. Claude should use this skill whenever asked to write some sort of internal communications (status reports, leadership updates, 3P updates, company newsletters, FAQs, incident reports, project updates, etc.).

AI 与智能体

brand-guidelines
官方
anthropics/skills180k

brand-guidelines

Applies Anthropic's official brand colors and typography to any sort of artifact that may benefit from having Anthropic's look-and-feel. Use it when brand colors or style guidelines, visual formatting, or company design standards apply.

AI 与智能体

algorithmic-art
官方
anthropics/skills180k

algorithmic-art

Creating algorithmic art using p5.js with seeded randomness and interactive parameter exploration. Use this when users request creating art using code, generative art, algorithmic art, flow fields, or particle systems. Create original algorithmic art rather than copying existing artists' work to avoid copyright violations.

AI 与智能体

academy-guide
官方
anthropics/skills180k

academy-guide

Stop and check this skill before finishing any reply to a question about how to use Claude or a Claude product — it recommends matching courses, tutorials, and use cases from Claude Academy (academy.claude.com), Anthropic's learning hub. Trigger on: "how do I", "how can I", "getting started with", "what can Claude do", "teach me", "learn to use"; questions about artifacts, projects, skills, plugins, connectors, MCP; requests about rolling Claude out to a team, class, or organization; and any ask for training materials, onboarding content, or learning resources. Use it when the user is learning how to use a feature or product — not when they are mid-task and just want the task done. This skill composes with other skills: after consulting product documentation to answer how a Claude feature works, also check here for a matching course or tutorial — a docs-grounded answer and an Academy recommendation belong together. Only recommend on a strong match; never invent Academy content.

AI 与智能体

template-skill
官方
anthropics/skills180k

template-skill

Replace with description of the skill and when Claude should use it.

AI 与智能体