Prompt-injection scanner for Chrome
AgentLens — See What Agents See
Reveal the instructions hiding behind the page.
See machine-readable webpage and PDF content that is not normally visible to you. AgentLens surfaces suspicious instructions, explains why they were flagged, and gives you control over what AI tools are allowed to read.
Platform
Google Chrome
Category
Developer Tools
Price
Free
See the instructions a person can miss—but an AI agent may still read.
Scan the current page or a supported PDF, inspect every suspicious signal with readable evidence, then reveal, strip, or allow the content before an AI tool encounters it.
Features
Detect instruction overrides, credential requests, data-exfiltration attempts, tool-use commands, concealment tactics, and common obfuscation patterns.
Reveal hidden instructions on the original page, highlight their source, and move between every detected signal.
Strip suspicious instruction sources before sharing page content with an AI tool, then restore the original page by reloading.
Review why each finding was flagged, how it was concealed, which extraction profiles can encounter it, and which rule matched.
Scan supported PDF text layers, annotations, metadata, clipped content, and off-page content without uploading the document.
Enable optional Live protection, keep a local scan history, customize deterministic rules, and export bounded reports on demand.
Version
0.1.0
Updated
August 23, 2026
Size
732 KiB
Processing
On-device
Accounts
None
Telemetry
None