Armor vs Everything
Side-by-side comparison of two Model Context Protocol servers. Pick the right one for Claude Desktop, Claude Code, or Cursor.
Drop-in Rust sidecar that wraps any MCP server: scans tool calls for prompt injection (UTS-39 confusable-fold + Unicode strip + Aho+Regex + directory-traversal)
Reference / test server with prompts, resources, and tools.
Comparison
| Feature | Armor | Everything |
|---|---|---|
| Pricing | Free | Free |
| Installs | — | — |
| Rating | — | — |
| Verified | — | |
| Hosted | — | — |
| Tools | — | — |
| Category | development | development |
| Author | studiomeyer-io | Community |
| Repo | studiomeyer-io/mcp-armor | modelcontextprotocol/servers |
When to pick Armor
Drop-in Rust sidecar that wraps any MCP server: scans tool calls for prompt injection (UTS-39 confusable-fold + Unicode strip + Aho+Regex + directory-traversal), validates Ed25519 manifest signatures (TOFU keystore + Sigstore Rekor bridge), strips loader-class env keys (LD_PRELOAD, NODE_OPTIONS, ...) from spawned children, detects tools/list schema drift (Layer 7: rug-pull / silent-redefinition, NFKC fingerprint, BLAKE3 or SHA-256 for FIPS/PCI/HIPAA, JCS), scans tool descriptions + full schemas for first-sight tool poisoning / full-schema poisoning (Layer 8, OWASP MCP03 — the poisoned-baseline gap drift can't see), serves a fully-wired rmcp 1.5 control-plane (MCP spec 2025-11-25, closes CVE-2026-42559 transitively), exports OTLP gRPC telemetry, blocks marketplace-poisoning vectors. <5ms p99 overhead.
When to pick Everything
Reference / test server with prompts, resources, and tools.
Looking for something else? Browse all MCPs or check trending this week.