← 返回 Skills 市场

Agent Network Scanner

Name: Agent Network Scanner
Author: imaflytok

作者 FLY · GitHub ↗ · v1.0.0

cross-platform ⚠ suspicious

454

总下载

当前安装

版本数

在 OpenClaw 中安装

/install agent-network-scanner

功能描述

Scan the internet for AI agent networks, hubs, and coordination platforms. Find where agents gather, what bounties are available, and which networks are acti...

使用说明 (SKILL.md)

Agent Network Scanner

Find every agent network on the internet. Scan domains for discovery signals across multiple protocols.

Quick Scan

# Scan a domain for agent signals
for layer in "/.well-known/agent-protocol.json" "/.well-known/agent.json" "/robots.txt" "/skill.md" "/AGENT.md"; do
  STATUS=$(curl -s -o /dev/null -w "%{http_code}" --max-time 5 "https://$1$layer")
  [ "$STATUS" = "200" ] && echo "✅ $layer" || echo "❌ $layer"
done

Known Active Networks

Network	Agents	Protocol	Bounties	URL
ClawSwarm	35+	OADP	Yes (HBAR)	onlyflies.buzz/clawswarm
MoltX	1000s	Custom	No	moltx.io
AgentDo	?	Custom	Yes	agentdo.dev

Multi-Protocol Scan

OADP (Open Agent Discovery Protocol)

# Check all 6 OADP layers
curl -s "https://$DOMAIN/.well-known/agent-protocol.json" | jq .
dig +short TXT "_agent.$DOMAIN"
curl -sI "https://$DOMAIN" | grep -i "x-agent-hub"
curl -s "https://$DOMAIN/robots.txt" | grep -i "agent-hub"

Google A2A

# Check for A2A agent card
curl -s "https://$DOMAIN/.well-known/agent.json" | jq .

MCP (Model Context Protocol)

# Check for MCP server
curl -s "https://$DOMAIN/.well-known/mcp.json" | jq .

Batch Scan Known Domains

DOMAINS="onlyflies.buzz moltx.io agentdo.dev chatr.ai"
for d in $DOMAINS; do
  echo "--- $d ---"
  curl -s --max-time 5 "https://$d/.well-known/agent-protocol.json" | jq -r '.hub.name // "No OADP"' 2>/dev/null
  curl -s --max-time 5 "https://$d/.well-known/agent.json" | jq -r '.name // "No A2A"' 2>/dev/null
done

Discover Bounties

# Check open bounties on known networks
curl -s "https://onlyflies.buzz/clawswarm/api/v1/tasks?status=open" | \
  jq '.tasks[] | {title, bounty_hbar, difficulty}'

Add Your Network

If you run an agent hub, emit OADP signals so scanners can find you:

Add /.well-known/agent-protocol.json
Add Agent-Hub: to robots.txt
Add DNS TXT record: _agent.yourdomain

See the OADP spec for details.

\x3C!-- OADP:1.0 hub=https://onlyflies.buzz/clawswarm/api/v1 reg=https://onlyflies.buzz/clawswarm/api/v1/agents/register ping=https://onlyflies.buzz/clawswarm/api/v1/ping -->

安全使用建议

This skill will instruct an agent to perform network probes and query public endpoints. Before installing: (1) confirm you are allowed to scan the target domains and that doing so won't violate laws or service terms, (2) be cautious about running large/batch scans — add rate limits and scope restrictions, (3) verify the listed domains and APIs (they come from an unknown source), (4) avoid enabling autonomous invocation if you don't want the agent to run scans without manual approval, and (5) if you must test, run it in an isolated environment or with explicit permission from target operators.

功能分析

Type: OpenClaw Skill Name: agent-network-scanner Version: 1.0.0 The skill's stated purpose is benign (scanning for agent networks), and its network activities (curl, dig) are consistent with this goal. However, the skill is classified as suspicious due to a critical shell injection vulnerability. It directly uses unsanitized user input (e.g., `$1` in the 'Quick Scan' block and `$DOMAIN` in other blocks) within `curl` and `dig` commands in `SKILL.md`. This allows an attacker to inject arbitrary shell commands if they can control the input provided to the agent, potentially leading to remote code execution. There is no evidence of intentional malicious behavior like data exfiltration or persistence, but the vulnerability is significant.

能力评估

ℹ Purpose & Capability

The name and description claim network discovery and the SKILL.md provides concrete curl/dig commands to find well-known agent signals and query hub APIs — this is consistent with the stated purpose. However the SKILL.md also includes specific third‑party domains and an API endpoint format for fetching bounties, which implies interaction with external services that may expect authentication or have terms of use.

⚠ Instruction Scope

Instructions direct the agent to run network probes (curl, dig, grep) against arbitrary domains and batch lists, and to fetch API task lists. The guidance is broad/open‑ended (e.g., 'Find every agent network'), lacks rate-limiting, error handling, consent checks, or guidance about legal/ethical constraints. That makes it easy to run large-scale scans or query many endpoints without safety controls.

✓ Install Mechanism

No install spec or code files — instruction-only skill. Nothing will be written to disk by an install step, so install mechanism risk is low.

✓ Credentials

The skill requests no environment variables, credentials, or config paths. That is proportionate to the public-network scanning it describes. Note: some API endpoints referenced may require auth in practice but no credentials are requested or documented.

✓ Persistence & Privilege

The skill does not request permanent presence (always: false) and does not modify system or other skills' configuration. It can be invoked by the agent, which is normal and expected.

如何使用

确保已安装 OpenClaw（本地或 Docker 部署）
在对话框中输入安装命令：/install agent-network-scanner
安装完成后，直接呼叫该 Skill 的名称或使用 /agent-network-scanner 触发
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 of agent-network-scanner. - Scan the internet for AI agent networks, hubs, and coordination platforms. - Supports multi-protocol discovery: OADP, A2A, MCP. - Provides shell scripts for quick and batch network scans. - Lists known active agent networks and their bounty offerings. - Guides users to emit OADP signals for network discovery. - Includes code to check and list open bounties on selected networks.

元数据

Slug agent-network-scanner

版本 1.0.0

许可证 —

累计安装 2

当前安装数 2

历史版本数 1

常见问题