360 Secure SkillHub
360 安全技能中心

Kaito

该 Skill 通过 Kaito MCP 工具调取实体或叙事的 mindshare 数据,按时间范围计算周均值、历史高低与平均、排名区间及叙事涨跌分类,并以文本或结构化数据呈现趋势;独立使用前会确认参数,也可嵌入工作流。

yh9277
yh9277
热度 81.37
最近更新 2026-05-11 17:29:59
分类:数据依赖:MCP 服务应用场景:市场分析 / 报告撰写 / 研究 / 社交媒体

安装技能

复制下面的话术发给你的 Agent,它会通过 secure-skills 检索并安装该技能。

执行命令安装此 skill: npx secure-skills-cli add kaito
Metadata
name
mindshare
description
Mindshare tracking primitive using Kaito MCP tools. Use this skill to check mindshare trends for an entity or narrative. Triggers on phrases like 'mindshare for SOL', 'how is the AI narrative trending', 'mindshare trend for Hyperliquid', 'which narratives are gaining share', 'mindshare comparison'. This is a primitive skill — it fetches mindshare data via kaito_mindshare, kaito_narrative_mindshare, or kaito_mindshare_delta and presents trend analysis. Can be called standalone or composed into workflow skills (social-listening, strategy, investigation). Do NOT use for sentiment (sentiment skill), content search (search skill), or user profiling (user-profile skill).

Mindshare Skill (Primitive)

Track mindshare trends for entities or narratives using Kaito MCP tools.

Parameters

  • entity (required): One of:
    • token: An entity/project (e.g. HYPE, SOL, ETH) — uses kaito_mindshare
    • narrative: A Kaito narrative ID (e.g. AI, DeFi, L2) — uses kaito_narrative_mindshare
  • time_range (optional): Period to analyze (default: 12 months for entities, flexible for narratives)
  • delta (optional): If true, also fetch kaito_mindshare_delta for recent changes

When called standalone, confirm parameters with the user first. When called by a workflow skill, parameters are passed directly.

Valid Narrative IDs (case-sensitive)

AI, DeFi, Stablecoin, L2, RWA, ZK, Meme, DePIN, SocialFi

Known invalid narrative IDs (use kaito_advanced_search with keyword as fallback): perp dex, prediction markets, privacy, restaking, NFT, gaming, liquid staking, robotics.


Workflow

For Entity Mindshare

Call kaito_mindshare with:

  • token: <TOKEN>
  • start_date: <range_start> (default: 12 months ago)
  • end_date: <today>

If all-zero results returned, switch from ticker to full project name (e.g. HYPE → HYPERLIQUID) and retry.

Optionally call kaito_mindshare_delta for a quick snapshot of recent change.

For Narrative Mindshare

Call kaito_narrative_mindshare with:

  • narrative: <NARRATIVE_ID> (case-sensitive)
  • start_date: <range_start>
  • end_date: <today>

Analysis

Present:

  • Trend chart — weekly averages over the time range, current level highlighted
  • Historical context — where current value sits vs high/low/average for the period
  • Rank interpretation (for entities):
    • Top 10 = dominant
    • Top 20 = strong
    • Top 50 = moderate
    • 50 = weak

  • Movement (for narratives):
    • % change = (last day - first day) / first day × 100
    • Surging: >= +10% relative change
    • Fading: <= -10% relative change
    • Stable: within ±10%

Output Format

When used standalone:

Mindshare: [ENTITY] | [TIME RANGE]

Trend: [weekly averages, current level]
Current: [value] ([rank interpretation if entity])
12m High: [value] | Low: [value] | Avg: [value]
Movement: [% change, surging/fading/stable]

When called by a workflow skill, return the structured data for the workflow to integrate.


Key Rules

  • If all-zero data for a ticker, switch to full name — Kaito indexes some entities by name rather than ticker. All-zero means the ticker isn't matched.
  • Narrative IDs are case-sensitive — use exactly: AI, DeFi, Stablecoin, L2, RWA, ZK, Meme, DePIN, SocialFi. Mismatched casing returns empty results silently.
  • Rank context is relative — Top 10 = dominant, Top 20 = strong, Top 50 = moderate, >50 = weak. Always provide this context so users can calibrate.
  • For narrative mindshare, calculate % change — raw numbers need the surging/fading/stable classification to be actionable.
  • Always present historical context — current value alone is meaningless without the 12-month high/low/average for comparison.