lw-firewall-rate-limit-worker
Pass
Audited by Gen Agent Trust Hub on Sep 9, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill serves as technical documentation for the LW Firewall plugin. It details the internal logic of the MU-plugin worker, including IP detection, endpoint classification, and storage backends. The provided PHP code snippet demonstrates standard integration for route-local protection.
- [EXTERNAL_DOWNLOADS]: The skill references the official GitHub repository for the plugin project (github.com/lwplugins/lw-firewall). This is a standard reference for software documentation and does not involve the execution of untrusted remote code.
- [INDIRECT_PROMPT_INJECTION]: The skill describes processing untrusted external data such as URI paths, query parameters, and cookie values for classification. This constitutes a potential attack surface if an agent processes malicious data through these fields. Evidence Chain: 1. Ingestion points: URI paths, query strings, and cookies (SKILL.md). 2. Boundary markers: Not explicitly defined for agent-side processing. 3. Capability inventory: Command execution is limited to internal plugin logic (RateLimiter, AutoBanner) and 429 response generation. 4. Sanitization: The skill mentions path decoding and sanitization in
lw_firewall_parse_uri()andlw_firewall_path_is().
Audit Metadata