← 返回
未分类 Key

agentpass

Control Home Assistant devices through the agentpass security gateway. Use when the user asks to control lights, switches, sensors, climate, or any smart hom...
通过 agentpass 安全网关控制 Home Assistant 智能家居设备,包括灯光、开关、传感器、气候控制等。
torbenwetter
未分类 clawhub v1.0.0 1 版本 100000 Key: 需要
★ 0
Stars
📥 342
下载
💾 0
安装
1
版本
#latest

概述

agentpass — Home Assistant Gateway

Execute Home Assistant actions through a secure approval gateway. Read-only queries execute instantly. State-changing actions (turning lights on/off, calling services) are sent to a human guardian on Telegram for approval before execution.

IMPORTANT: The agentpass request command is a BLOCKING command. For tools that need approval, it will wait up to 15 minutes for the guardian to respond on Telegram. You MUST use background: false when executing agentpass request commands to prevent auto-backgrounding. The command returns the actual execution result once approved, or an error if denied/timed out. Do NOT tell the user to "check Telegram" or "approve the request" — just wait silently for the command to finish and then report the result.

Commands

List available tools

agentpass tools

Execute a tool

agentpass request <tool_name> [key=value ...]

Output is JSON on stdout. Errors go to stderr.

Exit codes: 0 = success, 1 = denied, 2 = timeout, 3 = connection error, 4 = invalid args.

Available Tools

Read-only (auto-approved, instant)

Get a single entity state:

agentpass request ha_get_state entity_id=light.living_room

Get all entity states:

agentpass request ha_get_states

List available HA services and their fields:

agentpass request ha_get_services

Get state history for an entity (last 24h):

agentpass request ha_get_history entity_id=sensor.temperature

Get logbook entries for an entity (last 24h):

agentpass request ha_get_logbook entity_id=light.living_room

Get Home Assistant configuration:

agentpass request ha_get_config

Requires human approval (command blocks until resolved)

These commands block until approved or denied. Always use background: false to prevent auto-backgrounding.

Call a Home Assistant service:

# exec with background: false
agentpass request ha_call_service domain=light service=turn_on entity_id=light.living_room
agentpass request ha_call_service domain=light service=turn_on entity_id=light.bedroom brightness=128 color_name=red
agentpass request ha_call_service domain=switch service=toggle entity_id=switch.fan
agentpass request ha_call_service domain=climate service=set_temperature entity_id=climate.thermostat temperature=21

The domain and service args are used for the URL path. All other args (entity_id, brightness, color_name, temperature, etc.) are sent as the JSON request body.

Render a Home Assistant template:

# exec with background: false
agentpass request ha_render_template template="{{ states('sensor.temperature') }} degrees"

Always denied (blocked by policy)

  • ha_fire_event — firing custom events is blocked
  • ha_call_service with domain=lock — lock control is blocked

Important Notes

  • entity_id format: Always domain.object_id, e.g. light.living_room, sensor.temperature, switch.garden_pump. Must be lowercase with underscores.
  • domain/service format: Lowercase with underscores, e.g. light, turn_on, set_temperature.
  • Approval timeout: If the human guardian doesn't respond within 15 minutes, the request is automatically denied.
  • Discover entities first: If you don't know an entity's ID, run ha_get_states to list all entities, or ha_get_services to see available services and their parameters.
  • Service parameters: Use ha_get_services to discover which parameters a service accepts (e.g. brightness, color_name, rgb_color, temperature, hvac_mode).

版本历史

共 1 个版本

  • v1.0.0 当前
    2026-05-07 12:24 安全 安全

安全检测

腾讯云安全 (Keen)

安全,无风险
查看报告

腾讯云安全 (Sanbu)

安全,无风险
查看报告

🔗 相关推荐

ai-intelligence

Self-Improving + Proactive Agent

ivangdavila
自我反思+自我批评+自我学习+自组织记忆。智能体评估自身工作、发现错误并持续改进。
★ 1,362 📥 318,822
security-compliance

Skill Vetter

spclaudehome
AI智能体技能安全预审工具。安装ClawdHub、GitHub等来源技能前,检查风险信号、权限范围及可疑模式。
★ 1,218 📥 266,713
ai-intelligence

ontology

oswalpalash
类型化知识图谱,用于结构化智能体记忆与可组合技能。支持创建/查询实体(人员、项目、任务、事件、文档)及关联...
★ 714 📥 244,034