安全审计
- 作者仓库星标 0
- 作者更新于 实时读取
- 作者仓库 nano-core
- 领域
- AI 智能
- 兼容 Agent
-
- Claude Code
- Cursor
- Cline
- Codex
- Windsurf
- Gemini CLI
- +20
- 信任分
- 88 / 100 · 社区维护
- 作者 / 版本 / 许可
- @0-CYBERDYNE-SYSTEMS-0 · 未声明 license
- Token 消耗评级
- 低消耗
- 接入复杂程度
- 即装即用
- 是否需要外部 API Key
- 不需要
- 兼容的系统
- macOS · Linux · Windows
- 底层运行要求
- 无特殊要求
- 文件与系统权限
-
- 只读
- 允许写入 / 修改
- 网络行为
- 仅限本地
- 安装命令数
- 26 条
档案由构建时根据 SKILL.md 与安装命令自动衍生,可能与作者实际意图存在差异。
需要注意: 未限定 allowed-tools,默认拥有全部工具权限。
---
name: fft-demo-holodeck-ops
description: Drive the FFT_demo_dash holodeck system — scene triggers, agent commentary, canvas presets, and…
category: AI 智能
runtime: 无特殊运行时
---
# fft-demo-holodeck-ops 输出预览
## PART A: 任务判断
- 适用问题:提示词、Agent 工作流、模型评估或自动化推理。
- 输入要求:目标材料、限制条件、期望输出和验收方式。
- 证据边界:围绕“When to use this skill / When not to use this skill / Guardrails”读取原文规则,不把推断写成作者承诺。
## PART B: 执行结果
- **01** 任务判断:确认你的需求是否属于提示词、Agent 工作流、模型评估或自动化推理,并标出输入、限制和预期结果。
- **02** 执行计划:优先按“When to use this skill / When not to use this skill / Guardrails”拆成步骤,说明每一步会读取什么、修改什么、产出什么。
- **03** 交付结果:给出可复制的命令、文件改动、检查清单或内容草稿,并说明如何继续迭代。
- **04** 风险边界:结合 读取文件、写入/修改文件、主要在本地完成、通常不需要额外 API Key 给出执行前确认项。
## Running Rules
- 读取文件、写入/修改文件;主要在本地完成;通常不需要额外 API Key。
- 先小样例验证,再放大到真实任务。
- 交付时同时给结果、检查口径和下一步迭代建议。 原文出现了 `/workspace` 这类斜杠命令;如果你的 Agent 支持命令触发,优先用命令开场,再补充目标和边界。
告诉 Agent 目标文件或材料、期望结果、不可改范围、是否允许联网或执行命令。本 Skill 的权限画像是:读取文件、写入/修改文件。
先用一个小任务确认它会围绕“When to use this skill / When not to use this skill / Guardrails”工作;涉及文件或命令时,先看 diff、日志、预览或测试结果。
检查最终产物是否包含明确结果、必要证据和下一步动作;如果输出泛泛而谈,就补充输入、边界和验收标准后重跑。
---
name: fft-demo-holodeck-ops
description: Drive the FFT_demo_dash holodeck system — scene triggers, agent commentary, canvas presets, and…
category: AI 智能
source: 0-CYBERDYNE-SYSTEMS-0/nano-core
---
# fft-demo-holodeck-ops
## 什么时候使用
- fft-demo-holodeck-ops 是 AI 智能方向的技能,主要扩展 Agent 在调模型、改提示词、跑评测这类场景下的能力 适合处理AI Agent、提示词、模型评估与自动化推理,核心价值是把输入、判断、执行、验证和交付边界…
- 面向提示词、Agent 工作流、模型评估或自动化推理,优先处理能明确输入、步骤和验收标准的工作。
## 需要提供什么
- 目标材料、目录范围、期望结果和不可改动内容。
- 是否允许联网、执行命令、读写文件或调用外部服务。
## 执行规则
- 围绕「When to use this skill / When not to use this skill / Guardrails」组织步骤,不把推断写成作者事实。
- 读取文件、写入/修改文件;主要在本地完成;通常不需要额外 API Key。
- 先跑小样例,确认结果可检查后再扩大任务范围。
## 输出要求
- 给出最终产物、关键证据、验证方式和下一步动作。
- 信息不足时标记 unknown,不编造命令、平台或依赖。 作者原文负责流程事实;仓库文件负责来源和命令;流狐只补充适用场景、限制和质量判断。
skill "fft-demo-holodeck-ops" {
输入层 -> 用户目标 + 目标文件 + 禁止范围 + 验收标准
上下文层 -> When to use this skill / When not to use this skill / Guardrails
规则层 -> SKILL.md 触发条件 / 执行顺序 / 输出格式
运行层 -> 无特殊运行时 | 读取文件、写入/修改文件 | 主要在本地完成
安全层 -> 通常不需要额外 API Key + 小任务验证 + diff / 日志复核
输出层 -> 可复制结果 + 检查清单 + 下一步迭代
} FFT Demo Holodeck Ops
Use this skill whenever you respond to a farm query and the FFT_demo_dash dashboard is live. Your Telegram reply and the dashboard should update together — the audience sees both simultaneously.
When to use this skill
- Any time you answer a farm question and the dashboard is running at
http://localhost:8123. - When a user asks about conditions, alerts, status, or operational decisions.
- When a scene change would make the answer more visually compelling.
- When showing the dashboard to investors, growers, compliance officers, or press.
When not to use this skill
- Do not use for production HA deployments with real physical equipment.
- Do not trigger scenes without reading current farm state first.
Guardrails
- Never run destructive git commands unless explicitly requested.
- Preserve unrelated worktree changes.
- Keep write operations main-chat-only.
- Always read
current.jsonbefore setting scene entities.
Rule: Always Write Agent Commentary
On every response to a farm query, write two entities before sending your Telegram reply:
1. Agent status (short mode label, ≤128 chars):
{
"type": "farm_action",
"action": "ha_set_entity",
"params": {
"entityId": "input_text.agent_status",
"value": "Storm Response Active"
},
"requestId": "act_<ts>_status"
}
2. Agent commentary (your full reasoning/answer, ≤500 chars):
{
"type": "farm_action",
"action": "ha_set_entity",
"params": {
"entityId": "input_text.agent_commentary",
"value": "NWS shows 68mph gusts arriving in ~3h. Pre-charging battery to 95%. Securing greenhouse vents. Activating storm drainage pumps."
},
"requestId": "act_<ts>_commentary"
}
The commentary card on the dashboard renders these live with an animated border. The audience sees the agent reasoning in real time.
Scene System
Trigger scenes by setting input_select.demo_scene via ha_call_service:
{
"type": "farm_action",
"action": "ha_call_service",
"params": {
"domain": "input_select",
"service": "select_option",
"data": {
"entity_id": "input_select.demo_scene",
"option": "Golden Hour Harvest"
}
},
"requestId": "act_<ts>_scene"
}
Available Scenes
| Scene Name | When to Use |
|---|---|
Golden Hour Harvest |
User asks about yields, harvest timing, solar peak, profitable operations |
Morning Irrigation Pulse |
User asks about irrigation, water management, dawn operations |
Midnight Ops |
User asks about night security, overnight battery drain, off-hours ops |
IPM Zone Lockdown |
User mentions pests, spray schedules, quarantine, Zone B issues |
Cannabis Full Bloom |
User asks about grow room status, flower stage, environment setpoints |
Fertigation Drift Crisis |
User asks about EC/pH, nutrient dosing, runoff, feed issues |
Emergency Cascade |
User mentions storm + security together, asks about emergency protocols |
Drone Surveillance Sweep |
User asks about security sweep, perimeter check, anomaly detection |
VIP Tour |
User mentions investors, tour, showing the farm, "make it look good" |
Compliance Audit Crunch |
User mentions METRC, audits, state inspection, compliance scores |
Auto (Weather-Driven) |
Reset to live weather mode after a demo scene |
Each scene: updates theme, 8–20 entity values, writes its own agent_status/commentary automatically.
Canvas Presets
After triggering a scene, optionally load a matching canvas spec for the Agent Canvas view.
Canvas spec files are at /workspace/dashboard/canvas-specs/ (or the HA www path). Load via:
{
"type": "farm_action",
"action": "ha_canvas_set_spec",
"params": {
"spec": { "title": "...", "panels": [ ... ] }
},
"requestId": "act_<ts>_canvas"
}
Or read an existing preset first with ha_canvas_get_spec, then patch.
Live Weather Mirroring
The simulator is running live NWS weather for Cedar Creek TX (30.08, -97.49) by default.
To mirror a different city's live weather into a specific zone, use a Telegram-prompted bash call from the host (not IPC) — or tell the user:
npm run demo:mirror -- --location nyc --zone greenhouse_a
Available locations: cedar-creek, napa-valley, nyc, nyc-greenhouse, seattle, miami, denver, phoenix, florida-citrus
Available zones: greenhouse_a, greenhouse_b, greenhouse_c, outdoor_north, cannabis
Demo Flow: Farmer Ask → Dashboard Response
Every interaction should follow this sequence:
- Read
/workspace/farm-state/current.json— get live entity values and suggested theme - Decide which scene (if any) matches the query context
- Write
agent_status(mode label) - Write
agent_commentary(your answer in ≤500 chars) - Trigger scene if appropriate (via
input_select.demo_scene) - Reply via Telegram — same content as
agent_commentary, expanded if needed
The dashboard and Telegram reply should tell the same story simultaneously.
Example Interactions
"Storm coming tomorrow — what should I prepare?"
- Scene:
Emergency Cascade - Status:
Storm Prep Active - Commentary:
NWS advisory confirms 65mph gusts, 2.8" rain expected. Pre-charging battery to 95%. Securing vent actuators. Activating storm drainage. Harvest window closes in ~6h.
"I'm showing investors in 10 minutes"
- Scene:
VIP Tour - Status:
VIP Mode — All Systems Optimal - Commentary:
Solar at 380kW, battery 95%, all compliance at 98%+. Zero active alerts. Yields tracking 12% above projection. Farm is ready for inspection.
"Zone B IPM alert just came in"
- Scene:
IPM Zone Lockdown - Status:
IPM Protocol — Zone B Quarantine - Commentary:
Pest pressure Zone B: 7.5/10. Room B HVAC isolated to prevent spray drift. Drone launched for survey sweep. Azadirachtin queued. Days since last spray reset.
"How's the grow room doing right now?"
- Read current entities → if nominal, set
Cannabis Full Bloom - Status:
Grow Rooms Nominal — All Rooms Flowering - Commentary:
PPFD 920, DLI 48, air temp 78°F, RH 52%, CO₂ 1200ppm, EC 2.4, pH 6.0. All rooms in flower at optimal setpoints. Expected yield on track.
先判断是否适合
作者设计意图
作者的方法与取舍
边界和复核