The WAF for agents. Pattern-based + heuristic firewall scans prompts, RAG documents, tool argume...
84/100 A · ranked #246 of 339 MEOK MCPs · v1.0.6 · 5 tools
agent-prompt-injection-firewall-mcp scores 84/100 on the proofof.ai 100-point rubric — flagship-grade (top tier).
That is around the fleet average of 85. Install: pip install agent-prompt-injection-firewall-mcp.
| Category | Score | |
|---|---|---|
| README | 10/10 | |
| Tool design | 10/10 | |
| Examples | 4/10 | |
| Tests + CI | 10/10 | |
| Transports | 5/10 | |
| Metadata | 10/10 | |
| Reliability | 8/10 | |
| Security | 10/10 | |
| Docs surface | 7/10 | |
| Provenance/Revenue | 10/10 |
Framework: fastmcp · transport: stdio. Tools: scan_prompt, define_custom_rule, list_rules, scan_log, sign_firewall_attestation.
pip install agent-prompt-injection-firewall-mcpMCP client config: add {"command":"uvx","args":["agent-prompt-injection-firewall-mcp"]} under mcpServers.