Token导航 LogoToken导航TokenDH.com
研究检索敏感数据clawhub未标认证来源可访问clear审计通过

skytaleskytale 搜索

Agent Skill

skytale 用于查找、检索和筛选相关信息,适合在 OpenClaw 中需要根据关键词、任务场景或来源线索快速定位候选结果时使用。可结合来源仓库、安装命令和原始 README 继续核验具体用法。安装前建议确认权限范围、维护状态,以及是否会触发联网、命令执行或文件读写。

总安装

12,455

周安装

499

GitHub Stars

公开资料未说明

下载量

4,032
OpenClaw

安装说明

本站只整理中文说明和来源信息,不托管安装包,也不代用户安装。

GitHub

来源数

2

许可证

MIT-0

最后核验

2026-05-01

来源状态

来源可访问

安装方式

通过对话安装

复制提示词发给支持本地命令或 Skills 的 AI 助手,先确认命令和权限,再让它执行。

请帮我安装这个 Agent Skill:skytale(skytale 搜索)
来源仓库:https://github.com/nicholasraimbault/skytale
安装命令:
openclaw skills install skytale
安装前请先检查当前环境是否支持对应 CLI,并向我确认将要执行的命令、安装目录、联网范围和文件读写权限;确认后再执行。

命令行安装

复制命令到本机终端执行。该命令会通过 OpenClaw 从第三方来源获取 Skill;本站只展示命令,不托管安装包,也不自动执行。

ClawHubOpenClaw
openclaw skills install skytale

简介

AI Agent的信任层——加密通道、身份、审计、证明、信任圈、密钥轮换和联合。 MLS 协议(RFC 9420)。

SKILL.md

name
skytale
description
The trust layer for AI agents — encrypted channels, identity, audit, attestations, trust circles, key rotation, and federation. MLS protocol (RFC 9420).
version
0.5.1
metadata
openclaw
requires
env
anyBins
primaryEnv
SKYTALE_API_KEY
emoji
🔒
homepage
https://skytale.sh/docs/integrations/openclaw
os

Skytale Encrypted Shared Context

You have access to Skytale MCP tools for end-to-end encrypted agent messaging and shared context. All messages are encrypted using the MLS protocol (RFC 9420). The relay server cannot read message contents.

When to activate

Use Skytale tools when the user asks about:

  • Encrypted or secure communication between agents
  • Private messaging channels
  • Shared context, memory, or state between agents
  • Sending/receiving messages that must not be intercepted
  • Multi-agent coordination over encrypted channels
  • Agent identity, trust, or attestations
  • Key rotation or forward secrecy
  • Encrypted audit logging for compliance
  • Cross-organization agent federation
  • Anything mentioning "Skytale"

Prerequisites

The Skytale MCP server must be configured in your openclaw.json. If tools are unavailable, instruct the user to:

  1. Install: pip install skytale-sdk[mcp]
  2. Add the skytale MCP server to their openclaw.json (see examples/openclaw-config.json in the skill directory)
  3. Set SKYTALE_API_KEY environment variable (get one at https://app.skytale.sh)

Available MCP tools

Channel lifecycle

  • skytale_create_channel(channel) -- Create an encrypted channel. Channel names use org/namespace/service format (e.g. acme/research/results).
  • skytale_channels() -- List all active channels.

Messaging

  • skytale_send(channel, message) -- Send an E2E encrypted message to all channel members.
  • skytale_receive(channel, timeout) -- Receive buffered messages. Returns all messages since last check. Default timeout: 5 seconds.

Key exchange (manual)

  • skytale_key_package() -- Generate an MLS key package (hex-encoded). Used when manually adding members.
  • skytale_add_member(channel, key_package_hex) -- Add a member using their key package. Returns a hex-encoded MLS Welcome message.
  • skytale_join_channel(channel, welcome_hex) -- Join a channel using a Welcome message from the channel owner.

Multi-agent setup

For two agents to communicate:

  1. Agent A calls skytale_create_channel("org/team/channel").
  2. Agent B calls skytale_key_package() and shares the result with Agent A.
  3. Agent A calls skytale_add_member("org/team/channel", key_package_hex) and shares the Welcome with Agent B.
  4. Agent B calls skytale_join_channel("org/team/channel", welcome_hex).
  5. Both agents can now skytale_send and skytale_receive on the channel.

When using the hosted API with invite tokens (recommended), this handshake is automated -- the SDK handles key exchange through the API server.

SDK features (beyond MCP tools)

The Skytale SDK (0.5.1+) includes modules that agents can use alongside MCP tools for richer security capabilities. These are used via the Python SDK directly, not through MCP tool calls.

Encrypted Shared Context

Encrypted key-value state shared across all channel members. Supports typed entries, scoped access, TTL, and structured handoffs.

from skytale_sdk.context import SharedContext, ContextType

ctx = SharedContext(mgr, "acme/research/results")
ctx.set("task_status", {"phase": "analysis", "progress": 0.7})
ctx.set("search_results", results, type=ContextType.ARTIFACT)
ctx.set("private_analysis", data, visible_to=["did:key:z6MkAgentB..."])
ctx.handoff("did:key:z6MkAgentB...", {"task": "summarize"}, tried=["approach_1"])
ctx.subscribe(lambda key, val: ..., type_filter=ContextType.HANDOFF)

Agent Identity

Cryptographic Ed25519 DID:key identities for agents. Generate, sign, and verify. Supports DID:key and DID:web identity types.

from skytale_sdk import AgentIdentity
identity = AgentIdentity.generate()
print(identity.did)  # did:key:z6Mk...
signature = identity.sign(b"message")

Key Rotation

Rotate MLS leaf keys for forward secrecy. Generates a new UpdatePath commit and advances the group epoch.

mgr.rotate_key("acme/secure/channel")

Encrypted Audit Logging

Hash-chained tamper-evident logs encrypted with MLS exporter secrets. Stored server-side as opaque ciphertext for compliance (EU AI Act Article 12).

from skytale_sdk import SkytaleChannelManager
mgr = SkytaleChannelManager(identity=b"agent", api_key="sk_live_...", audit=True)
audit_key = mgr.export_audit_key("acme/secure/channel")

from skytale_sdk.audit import EncryptedAuditLog
enc_log = EncryptedAuditLog(mgr.audit_log, audit_key)
enc_log.record({"event": "analysis_complete"})
encrypted_entries = enc_log.pending_entries()

Cross-Organization Federation

Join channels across organizations using federation invite tokens.

mgr.join_federation("partner-org/shared/channel", "skt_fed_abc123...")

Attestations

SD-JWT attestations for agent trust and reputation. Issue claims about other agents with selective disclosure.

from skytale_sdk import AgentIdentity, create_attestation, verify_attestation
issuer = AgentIdentity.generate()
att = create_attestation(issuer, "did:key:z6Mk...", {"task_score": 0.95}, disclosed=["task_score"])
valid = verify_attestation(att, issuer.public_key)

Trust Circles

Credential-gated MLS groups. Only agents meeting admission policies can join.

from skytale_sdk import TrustCircle, AdmissionPolicy
policy = AdmissionPolicy(required_claims=["certification"], min_score=0.8)
circle = TrustCircle.create(mgr, "acme/trusted/group", policy)

Agent Registry

Discover agents by capability via the Skytale API. Supports visibility tiers (public, organization, private).

mgr.register_agent(capabilities=["analysis", "summarization"])
agents = mgr.search_agents(capability="analysis")

Rules

  • NEVER log, display, or include encryption keys, key packages, or Welcome messages in user-visible output. Treat them as opaque tokens passed between tools only.
  • NEVER include API keys in messages sent through channels.
  • Channel names MUST follow org/namespace/service format.
  • Always call skytale_receive with a reasonable timeout (2-10 seconds). Do not poll in tight loops.
  • When creating channels for the user, suggest descriptive names matching their use case.
  • If skytale_receive returns no messages, inform the user and offer to check again -- do not retry silently.

Error handling

  • MCP tools not available: Tell the user to configure the Skytale MCP server in their openclaw.json and install skytale-sdk[mcp].
  • Authentication failure: Check that SKYTALE_API_KEY is set and valid. Direct the user to https://app.skytale.sh to obtain a key.
  • Channel not found on receive/send: The channel must be created or joined first.
  • Listener died: The background connection was lost. Recreate the channel.

适合场景

01

OpenClaw 用户查找和安装 Skill 时

02

用户想查找某类 Agent Skill 时

03

需要根据任务场景推荐可安装能力包时

04

需要对比不同来源的安装命令和来源信息时

能力概览

能力 1

按任务关键词查找相关 Skills

能力 2

展示可复制的安装命令

能力 3

保留来源站点、仓库和原始说明,方便继续核验

能力 4

补充不同宿主或平台的使用分布数据

能力 5

展示第三方安全扫描或审计结果

安装后应在对应宿主中按原始 README 的触发条件使用;具体调用方式请以来源页面和 README 为准。

平台分布

OpenClaw

82.82%
按下载量换算3,339

安全审计

VirusTotal

通过

ClawScan

通过

Static analysis

通过

权限和风险

敏感数据

该 Skill 可能接触密钥、Token、环境变量或敏感配置,应进入高风险复核队列,默认不自动发布。

安装前确认

本站仅展示第三方公开信息,不托管安装包,不提供自动安装或运行环境。安装前应自行审查源码、依赖和命令行为。当前只有一个来源,正式发布前建议补源仓库或其他目录站核验。

来源信息

继续浏览同类 Skills