← 返回 Skills 市场
batsirai

aeo-system

作者 Batsirai Chada · GitHub ↗ · v1.0.0
cross-platform ⚠ suspicious
439
总下载
0
收藏
1
当前安装
1
版本数
在 OpenClaw 中安装
/install aeo-system
功能描述
Answer Engine Optimization — get AI assistants to recommend your brand. Run AEO audits, build Answer Intent Maps, track AI recommendation positions, and main...
使用说明 (SKILL.md)

AEO (Answer Engine Optimization) System

Get AI assistants — ChatGPT, Perplexity, Claude, Gemini — to recommend your brand when people ask purchase-intent questions.


What This Is

AEO is the discipline of optimizing for AI-powered answer engines the same way SEO optimizes for search engines. When someone asks Perplexity "what's the best magnesium supplement for sleep?" — AEO determines whether your brand gets named.

This skill gives an OpenClaw agent the ability to:

  1. Audit a brand's current AEO infrastructure across all 7 layers
  2. Map which brands AI platforms recommend (and in what position) for any category
  3. Track position changes week over week
  4. Build the missing infrastructure (Answer Hub, brand-facts.json, schema, citations)
  5. Maintain the system with a weekly 90-minute protocol

When to Load This Skill

  • User asks to "run an AEO audit" for a brand or URL
  • User asks "which brands are being recommended by AI for [category]?"
  • User asks to "build an Answer Intent Map" for a category
  • User asks to check a brand's Answer Hub, brand-facts.json, or schema markup
  • User asks to track AI recommendation positions over time
  • User asks to run the "weekly AEO maintenance protocol"

The 7-Layer AEO Framework

Layer Name What It Is Priority
1 Answer Intent Map Spreadsheet of all purchase-intent queries + which brands AI recommends Foundation
2 Answer Hub A long-form guide page that answers every key question in your category High
3 Brand-Facts Page Human-readable brand facts page (neutral, factual, cite-able) High
4 brand-facts.json Machine-readable brand data at /.well-known/brand-facts.json Medium
5 Schema Markup Product, FAQ, and Organization structured data Medium
6 Citation Network Getting listed on the sources AI models actually cite High
7 GPT Shopping Google Merchant Center + review feed for AI shopping results High

Prerequisites

Required:

  • PERPLEXITY_API_KEY — enables direct API queries (get free at perplexity.ai/settings/api)
  • Node.js v18+ (for the answer-intent-map.js script)

Optional:

  • OPENAI_API_KEY — enables ChatGPT query automation
  • BRAVE_API_KEY — enables web searches for infrastructure checks

Without API keys: The skill runs in "manual-assist" mode — generates the queries, provides a blank log template, and analyzes results you paste in.


Core Workflows

Workflow 1: Full AEO Audit

Trigger: "Run an AEO audit for [brand URL]"

Steps:

  1. Fetch and analyze the brand's website for AEO infrastructure:

    • Check for Answer Hub page (/guides/ or similar long-form page)
    • Check for Brand-Facts page (/brand-facts)
    • Check for machine-readable data (/.well-known/brand-facts.json)
    • Audit schema markup on product pages (via Rich Results API or web_fetch)
    • Check for a Wikidata entry
    • Check Google Merchant Center eligibility signals
  2. Score each of the 7 layers (0–3 scale):

    • 0 = Doesn't exist
    • 1 = Exists but incomplete or outdated
    • 2 = Exists, functional, minor gaps
    • 3 = Complete, current, optimized
  3. Generate a gap analysis report with:

    • Current score per layer
    • Priority order for implementation
    • Specific action items for each missing layer

Output: Markdown report saved as aeo-audit-[brand]-[date].md


Workflow 2: Answer Intent Map

Trigger: "Build an Answer Intent Map for [category]" or run scripts/answer-intent-map.js

Steps:

  1. Generate query list from four types:

    • Category queries: "best [product] for [use case]" (10–15 queries)
    • Comparison queries: "[brand] vs [competitor]" (10 queries)
    • Brand queries: "is [brand] worth it" (5 queries)
    • Educational queries: "does [ingredient] help with [condition]" (10 queries)
  2. For each query, query available platforms:

    • Perplexity API (structured JSON response with citations)
    • OpenAI API (text response — brand names extracted by parser)
    • Browser fallback for Claude and Gemini
  3. Parse responses to extract:

    • Brand names mentioned (position 1, 2, 3)
    • Source URLs cited
    • Key verbatim quotes
  4. Write results to JSON data file + Markdown summary report

Output: answer-intent-map-[category]-[date].json + .md summary

Run the script:

node scripts/answer-intent-map.js \
  --category "magnesium supplements" \
  --brand "MyBrand" \
  --queries 20

# Or with a config file:
node scripts/answer-intent-map.js --config ./aeo-config.json

Workflow 3: Weekly Maintenance Protocol

Trigger: "Run weekly AEO maintenance" or scheduled cron

Steps:

  1. Load the brand's Answer Intent Map (top 15 priority queries)
  2. Query ChatGPT and Perplexity for each priority query in fresh sessions
  3. Compare results against previous week's log (detect position changes)
  4. Generate maintenance report:
    • Position changes (up/down/new competitors)
    • New sources being cited this week
    • Recommended Answer Hub updates
  5. Check brand-facts.json for stale lastUpdated timestamp
  6. Check Google Merchant Center for disapprovals (via browser if needed)

Output: aeo-weekly-report-[date].md

Use the checklist: templates/weekly-maintenance-checklist.md


Workflow 4: Citation Network Analysis

Trigger: "Analyze AEO citations for [category]"

Steps:

  1. Run 20 category queries via Perplexity API (citations returned directly)
  2. Extract all unique source URLs from responses
  3. Group and count by domain
  4. Identify top 10 most-cited external sources in the category
  5. Generate outreach priority list

Output: Citation analysis report with target sites ranked by citation frequency


Workflow 5: Infrastructure Build

Trigger: "Build AEO infrastructure for [brand]" or "Set up brand-facts.json"

Steps:

  1. Ask for brand details (or load from aeo-config.json)
  2. Generate from templates:
    • brand-facts.jsontemplates/brand-facts.json (fill placeholders)
    • Answer Hub page → templates/answer-hub-template.md
    • Schema markup snippet (JSON-LD for product pages)
  3. Provide implementation instructions per asset

Configuration

Create aeo-config.json in your working directory:

{
  "brandName": "Your Brand Name",
  "brandUrl": "https://yourbrand.com",
  "category": "Magnesium Supplements",
  "priorityQueries": [
    "best magnesium supplement for sleep",
    "best magnesium glycinate supplement",
    "magnesium supplement for anxiety"
  ],
  "competitors": [
    "Competitor Brand A",
    "Competitor Brand B",
    "Competitor Brand C"
  ],
  "answerHubUrl": "https://yourbrand.com/guides/best-magnesium-supplements-2026",
  "brandFactsJsonUrl": "https://yourbrand.com/.well-known/brand-facts.json"
}

Output Files

File Description
aeo-audit-[brand]-[date].md Infrastructure audit report
answer-intent-map-[category]-[date].json Raw AI query results
answer-intent-map-[category]-[date].md Human-readable competitive summary
aeo-weekly-report-[date].md Weekly maintenance report
citation-analysis-[category]-[date].md Citation network analysis

Usage Examples

# Full infrastructure audit
"Run an AEO audit for mybrand.com"

# Build competitive intelligence
"Build an Answer Intent Map for the magnesium supplement category"

# Quick position check
"Check if mybrand.com is being recommended by Perplexity for 'best magnesium for sleep'"

# Weekly maintenance
"Run the weekly AEO maintenance protocol for my brand"

# Citation analysis
"Which sources does Perplexity cite most for collagen supplement recommendations?"

# Generate brand-facts.json
"Generate the brand-facts.json template for [brand details]"

# Scoring review
"Score my AEO infrastructure for mybrand.com on all 7 layers"

Platform Limitations

Platform Access Notes
Perplexity API (structured, reliable) Returns citations directly
ChatGPT API via OpenAI Text parsing required for brand extraction
Claude Browser required Generate queries + blank log; agent uses browser
Gemini Browser required Generate queries + blank log; agent uses browser

For Claude/Gemini: the skill generates the query list and a blank log template; use the browser tool to collect results.

Rate limits: Perplexity free tier ≈ 20 requests/minute. For 50+ queries, add --delay 3000 to the script.


File Structure

aeo-system/
├── SKILL.md                              ← This file
├── README.md                             ← Human-readable overview
├── scripts/
│   └── answer-intent-map.js              ← Core query automation script
└── templates/
    ├── answer-hub-template.md            ← Answer Hub page template
    ├── brand-facts.json                  ← Machine-readable brand data template
    └── weekly-maintenance-checklist.md  ← 90-minute weekly protocol

AEO System v1.0 — February 2026 A product by Carson Jarvis (@CarsonJarvisAI)

安全使用建议
This skill appears to implement what it claims (AEO audits, answer-intent maps) and its Node script calls Perplexity and OpenAI endpoints and writes local reports. Before installing or running it: 1) Do not assume the registry metadata is authoritative — SKILL.md and the script expect PERPLEXITY_API_KEY (required) and optionally OPENAI_API_KEY; the registry not listing env vars is an inconsistency to clarify with the author. 2) Only provide API keys scoped appropriately (use limited/rotating keys where possible); do not give Google Merchant Center, GA4, or other platform credentials to the skill as an automated secret—those steps are documented as manual in SKILL.md. 3) Review scripts/answer-intent-map.js yourself (or have a developer do so) to confirm there are no additional remote endpoints; the script uses api.perplexity.ai and api.openai.com only. 4) Run the script in an isolated environment (container/VM) if you plan to use your real API keys, and consider rate limits/costs for queries. 5) Ask the publisher to fix registry metadata to explicitly declare required environment variables and document what BRAVE_API_KEY refers to. If you need help verifying the code or limiting exposure, get a developer to audit the script before use.
功能分析
Type: OpenClaw Skill Name: aeo-system Version: 1.0.0 The OpenClaw AgentSkills skill bundle for 'aeo-system' is classified as benign. The `SKILL.md` clearly declares required environment variables and permissions (network, filesystem) which are consistent with the skill's stated purpose of Answer Engine Optimization. The `scripts/answer-intent-map.js` Node.js script uses only built-in modules, accesses API keys from environment variables, makes legitimate API calls to Perplexity and OpenAI, and performs file I/O (reading config, writing reports) within the declared permissions. There is no evidence of data exfiltration, malicious execution (e.g., `curl|bash`, `eval`), persistence mechanisms (e.g., explicit cron job creation instructions), obfuscation, or prompt injection attempts designed to subvert the agent's intended behavior or access unrelated sensitive data. All instructions and code align with the described AEO workflows.
能力评估
Purpose & Capability
The skill's stated purpose (AEO audits, intent maps, tracking AI recommendations) legitimately requires querying AI APIs and writing reports, which the included Node script and templates implement. However, the registry metadata claims 'Required env vars: none' while SKILL.md and the script require/expect PERPLEXITY_API_KEY (required), OPENAI_API_KEY (optional), and recommend BRAVE_API_KEY. That registry vs runtime mismatch is an incoherence that could mislead users about what secrets they'll need to provide.
Instruction Scope
SKILL.md instructs the agent to query external AI services (Perplexity and OpenAI), fetch/analyze brand websites, run schema/Rich Results checks, and produce files in the working directory. The included script performs HTTPS requests to api.perplexity.ai and api.openai.com and writes JSON/Markdown outputs. The instructions also reference manual checks (Google Merchant Center, GA4) and a 'browser fallback' for Claude/Gemini which are vague and leave room for human interaction or ad-hoc scraping. There are no instructions to read unrelated system files or exfiltrate environment variables; network and filesystem use is explicit.
Install Mechanism
No install spec — instruction-only skill with a Node.js script. This has a low install risk because nothing is downloaded or executed automatically beyond the included script. The script is plain JS using https/fs; no obfuscated or remote install steps were found.
Credentials
Requiring PERPLEXITY_API_KEY and optionally OPENAI_API_KEY is proportionate to querying those services. BRAVE_API_KEY is plausible for web-based checks but is less standard (clarify what 'BRAVE_API_KEY' refers to). The concern is the mismatch between the registry (no env vars declared) and SKILL.md/script (which require keys). That mismatch reduces transparency about what credentials will be requested. No unrelated or excessive environment variables are requested by the code itself.
Persistence & Privilege
The skill is not forced-always (always:false) and does not request elevated system or cross-skill privileges. It writes audit reports and data to the working directory (normal for this functionality) and does not modify other skills or global agent configuration.
如何使用
  1. 确保已安装 OpenClaw(本地或 Docker 部署)
  2. 在对话框中输入安装命令:/install aeo-system
  3. 安装完成后,直接呼叫该 Skill 的名称或使用 /aeo-system 触发
  4. 根据 Skill 的参数说明提供必要输入,即可获得结构化输出
版本历史
v1.0.0
Initial release
元数据
Slug aeo-system
版本 1.0.0
许可证
累计安装 1
当前安装数 1
历史版本数 1
常见问题

aeo-system 是什么?

Answer Engine Optimization — get AI assistants to recommend your brand. Run AEO audits, build Answer Intent Maps, track AI recommendation positions, and main... 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件,目前累计下载 439 次。

如何安装 aeo-system?

在 OpenClaw 或 Claude Code 对话框中运行命令「/install aeo-system」即可一键安装,无需额外配置。

aeo-system 是免费的吗?

是的,aeo-system 完全免费(开源免费),可自由下载、安装和使用。

aeo-system 支持哪些平台?

aeo-system 跨平台运行,可在任意部署了 OpenClaw / Claude Code 的环境中使用(cross-platform)。

谁开发了 aeo-system?

由 Batsirai Chada(@batsirai)开发并维护,当前版本 v1.0.0。

💬 留言讨论