Enter an X @username of an AI agent. We'll look up its profile, find its website, and run a heuristic ACP-SEC security analysis.
1
Lookup
2
Confirm
3
Results
🔍 Look up an Agent
Enter the X (Twitter) username of the AI agent you want to scan.
Example: @bankrbot, @virtuals_io, @aixbt_agent
✅ Confirm Agent Details
Review the scraped profile info and confirm the website URL to scan.
🤖
—
@—
Paste the agent's website, landing page, or docs URL. We'll scrape it for security signals.
If provided, we'll query Base mainnet for ACP Core registration
(0x238E…32E0) and surface an
ACP REGISTERED ✅ badge on the results.
🎯
You entered:
Will scan:
🤷 Don't have a website?
Try one of these public sources instead:
Parent organization website (e.g. anthropic.com for Claude)
GitHub repository (e.g. github.com/<org>/<agent>)
LinkedIn company page (linkedin.com/company/<name>)
Documentation site (often docs.<name>.com)
ℹ️
The scan analyzes publicly visible content and HTTP headers to infer
security posture. It does not run live agent probes —
for that, use acpsec check with an API key.
—
⚠️
No dedicated website found for this agent
Security score is capped at 15 / 100 in limited-scan mode.
Only AUTH-01 receives partial credit (agent identity declared).
All other dimensions require a public website to assess.
Provide the agent's website URL for a full 38-check scan.
0/ 100
🤖
—
—
⚡Base MCP Partner
⚠️Limited Scan
✅ACP Registered
—
Final Score
—
Critical Fails
—
Security Headers
30
Checks Run
HTTP Security Headers
—
Score Radar
Detailed Findings
Click a dimension to expand checks · Scores are inferred from public content
⚠ Scores are heuristic estimates based on public website content and HTTP headers.
Run acpsec check with a live agent endpoint for authoritative results.
