reflex-browser@0.15.0
Token-efficient browser automation for AI agents. Local MCP server + CLI: compressed page snapshots, batched actions, forensic failures.
AI Security Review
scanned 2d ago · by lpm-firewall-aiThe flagged credential and command activity maps to Reflex login, usage metering, browser profile management, and process checks used by the browser tool. The sensitive Chrome profile seeding is documented and supports the package’s stated browser automation purpose; no malicious data flow was identified.
Decision evidence
public snapshot- The manifest describes local browser automation and declares CLI binaries, with no install lifecycle hook.
- Login exchanges a one-time code with the Reflex login server and saves the returned Reflex API key locally.
- Usage reports go to the Reflex API and contain call counts, tool names, failure categories, platform, and response character counts.
- The README discloses persistent browser profiles and seeding from Chrome profiles as part of keeping sites signed in across runs.
- The package copies browser login data into its local automation profile by default on Windows and Linux, which is a sensitive capability disclosed in the README.
Source & flagged code
7 flaggedA single source file combines environment access, network access, and code or shell execution; review context before blocking.
dist/reflex.jsFindings
Affected versions and remediation
This report applies to reflex-browser@0.15.0.
See all versions of reflex-browser for other recorded verdicts.
- Review the evidence and your use of reflex-browser@0.15.0 before allowing it. Restrict the permissions described in this report.
- Choose an independently verified alternative or release. This report does not establish that other versions are safe.
Evidence last updated: .
Think this verdict is wrong? Dispute it on GitHub or read how reports are produced.