State of MCP security — October 2026
The 500 most-downloaded MCP servers (npm/PyPI) out of 23,902 official registry entries. Public data only; heuristic, not an audit.
- Unique advisories
- 112
- Current release affected
- 1
- Median score
- 93
- Archived / stale >1y
- 2/1
OSV.dev, across 6 servers (1%)
servers (0%) · 80 advisories
of 500 scored servers (mean 92)
of 476 with GitHub data
Score distribution
85–100 (strong)
475
70–84 (good)
24
50–69 (fair)
1
0–49 (poor)
0
Advisories by severity
| HIGH | 65 |
| CRITICAL | 30 |
| MODERATE | 17 |
Most-affected servers
| Server | Advisories | Affecting current | Score |
|---|---|---|---|
| PraisonAI | 104 | 80 | 56 |
| io.github.SepineTam/stata-mcp | 3 | 0 | 88 |
| Ouroboros | 2 | 0 | 94 |
| com.browser-use/browser-use | 1 | 0 | 96 |
| Excel MCP Server | 1 | 0 | 96 |
| io.github.CursorTouch/Windows-MCP | 1 | 0 | 96 |
Advisories published in October
- GHSA-2gpf-2492-q9jhHIGH PraisonAI — PraisonAI: Call API localhost-only authentication bypass via spoofed Host header
- GHSA-4r3p-w3mc-5v34MODERATE PraisonAI — PraisonAI: Prompt-injection defense blocks only when 3+ detector families fire simultaneously; realistic single-vector injections pass through unblocked
- GHSA-4w49-gwv8-fpjgHIGH PraisonAI — PraisonAI: Jobs API is unauthenticated by default and allows attacker-controlled webhook SSRF
- GHSA-5r6c-gj4g-r697HIGH PraisonAI — PraisonAI: SecurityPolicy command/path/import restrictions are completely unenforced by the default SubprocessSandbox backend
- GHSA-79fv-7hq9-w7xgHIGH PraisonAI — PraisonAI: API deploy code generator embeds unescaped YAML fields into Python source
- GHSA-7c92-x8vg-4258HIGH PraisonAI — PraisonAI: AgentMail webhook mode accepts forged unsigned message.received events and invokes agents
- GHSA-9mp3-24cc-77mgCRITICAL PraisonAI — PraisonAI: AICoder Arbitrary File Write and Command Execution via LLM Tool Calls
- GHSA-cv3g-hj65-pcfhHIGH PraisonAI — PraisonAI: Shell command allowlist bypass via find -exec built-in action
- GHSA-fj8f-m44g-c479HIGH PraisonAI — PraisonAI: Prompt Injection Defense Bypassed for HIGH-Level Threats
- GHSA-g6j7-pffp-8whgCRITICAL PraisonAI — PraisonAI: Code Injection via f-string Interpolation in Deploy API Server Generation
- GHSA-hc5v-gxvj-58whHIGH PraisonAI — PraisonAI: MCP HTTP-stream transport is unauthenticated by default, exposing tool enumeration and an unvalidated tool-call surface
- GHSA-q7m5-3jmv-vm48MODERATE PraisonAI — PraisonAI: ContextGatherer include resolution permits absolute and traversal reads outside the workspace
- GHSA-qj9c-59p6-8cgxHIGH PraisonAI — PraisonAI: AgentMail webhook lacks signature verification, allowing unauthenticated message injection and sender spoofing
- GHSA-wf65-4jjx-q444MODERATE PraisonAI — PraisonAI: PGVector and Cassandra knowledge stores interpolate vector dimensions into DDL
- GHSA-xpx6-x8c2-mw5wMODERATE PraisonAI — PraisonAI: Project custom command templates can read outside-workspace files into model prompts
Notable changes
- 2026-10-18 Archived OpenAdapt: your agent decides, OpenAdapt enters it and checks it saved — Repository OpenAdaptAI/openadapt-agent was archived
- 2026-10-18 Score drop OpenAdapt: your agent decides, OpenAdapt enters it and checks it saved — Score dropped 93 → 73
2 alert(s) this month. See all alerts.
Licenses
MIT (336) · Apache-2.0 (81) · none detected (34) · AGPL-3.0 (20) · no repo linked (19) · unknown (5) · BSD-3-Clause (2) · GPL-3.0 (1) · ISC (1) · CC-BY-SA-4.0 (1)
Frozen snapshot generated Oct 20, 2026 by scripts/report.mjs. Methodology · All reports