菜单
Advanced web research — search, extract, and synthesize information from multiple sources. Use when the user needs research, fact-checking, competitive analysis, or information gathering.
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
Access and search Telegram message history using AI. Search, summarize, extract action items, analyze sentiment, and recall conversations across all synced chats. Includes preset and custom AI skills, scheduled automations with Telegram bot delivery, and MCP integration.
ReAct (Reasoning + Acting) framework — interleave chain-of-thought reasoning with tool actions to solve complex tasks. Eliminates hallucination by grounding every reasoning step in real observations. Use when the agent needs multi-step reasoning with external tool use: question answering, fact verification, web navigation, interactive decision making, or any task where pure reasoning hallucinates and pure acting fails to plan.
Agent-to-agent feedback protocol for HermesHub. Register your agent identity, submit structured skill reviews with proof-of-use, and build trust scores for the skill ecosystem.
A meta-skill that silently watches your workflows and automatically generates reusable Hermes skills from them.
Comprehensive security hardening for Hermes Agent. Detects prompt injection, unicode smuggling, hidden directives, supply-chain skill poisoning, credential exposure, and memory manipulation. Provides runtime scanning rules, input/output validation patterns, and a defense-in-depth checklist aligned with OWASP LLM Top 10.
Undetectable, adaptive, high-performance Python web data extraction. Automatically survives website structure changes, bypasses anti-bot systems (Cloudflare, WAFs), and outperforms BeautifulSoup/Scrapy. Includes stealth browser fetching, CSS/XPath selectors, CLI, interactive shell, and MCP AI server integration.
| name | web-researcher |
| description | Advanced web research — search, extract, and synthesize information from multiple sources. Use when the user needs research, fact-checking, competitive analysis, or information gathering. |
| version | 1.0.0 |
| license | MIT |
| compatibility | Works with all Hermes backends. Optional TAVILY_API_KEY for enhanced results. |
| metadata | {"author":"hermeshub","hermes":{"tags":["research","web-search","extraction","summarization"],"category":"research","fallback_for_toolsets":["web"]}} |
| required_environment_variables | [{"name":"TAVILY_API_KEY","prompt":"Tavily API key (optional, enhances search quality)","help":"Get a free key at https://tavily.com","required_for":"enhanced search quality"}] |
Multi-source research agent with structured synthesis.
# Research: [Topic]
## Key Findings
- Finding 1 (Source: [url])
- Finding 2 (Source: [url])
## Detailed Analysis
[Structured analysis with inline citations]
## Sources
1. [Title](url) - Relevance: High/Medium/Low
2. [Title](url) - Relevance: High/Medium/Low
## Confidence & Gaps
- Confidence: High/Medium/Low
- Information gaps: [what couldn't be verified]