MCP Observatory
1m ago Tracked 58,273 Last 1m ago 24h 2,653
github TypeScript analyzed 32fb471

Jovancoding/Network-AI

github

Traffic light for AI Agents and TypeScript/Node multi-agent orchestrator with shared state, guardrails, and adapters for 32 AI frameworks

maintainer
Jovancoding
licence
MIT
first seen
2026-06-08
last seen
2026-09-15
releases · 30d
0
short id
risk 82/100 · heuristic grade
E critical inferred analyzer v33 (current)
  • vulnerabilities attested + 50
  • capability exposure inferred + 35
  • tool safety inferred + 5
  • trust mitigators mixed − 8

attested inferred mixed

The A–E grade is our heuristic synthesis — a "review this" prompt, not a verdict. Each factor is tagged by what backs it: attested (a verifiable record), reported (a third party's claim), or inferred (our own heuristic, e.g. permissions). See methodology.

grade last moved 1w ago · see ecosystem CVEs →

risk trajectory 6 movements
  • B · 32 → E · 82
  • C · 37 → B · 32
  • C · 57 → C · 37
  • C · 49 → C · 57
  • B · 32 → C · 49
  • A · 0 → B · 32
capability exposure grade factor +35
Inferred surface — each links to servers holding it:
vulnerabilities 13 CVEs · grade factor +50
CRITICAL
Improper Neutralization of Special Elements used in an OS Command in network-ai EPSS 1% CVE-2026-54051 affects ["< 5.9.1"]
CRITICAL
Network-AI: CVE-2026-46701 fix incomplete — empty default secret still authorizes all requests EPSS 0% CVE-2026-48814 affects ["<= 5.7.1"]
HIGH
Missing authentication on MCP HTTP endpoint allows unauthenticated privileged tool calls EPSS 0% CVE-2026-42856 affects ["<= 5.1.2"]
HIGH
SandboxPolicy blocklist and approval-gate bypass via quote/whitespace mismatch between matchers and… EPSS 0% CVE-2026-73615 affects ["<= 5.15.0"]
HIGH
ClaudeHookBridge deny-pattern gate bypass via 500-char extractTarget truncation before security dec… EPSS 0% CVE-2026-73614 affects ["<= 5.15.0"]
HIGH
Unauthenticated Cross-Origin MCP Tool Invocation via Empty Default Secret EPSS 0% CVE-2026-46701 affects ["<= 5.4.4"]
HIGH
Poisoned environment backup manifest allows arbitrary recursive deletion during backup pruning EPSS 0% CVE-2026-58484 affects ["<= 5.12.1"]
HIGH
APSAdapter Default Local Verifier Accepts Any Non-Empty Signature affects ["<= 5.13.3"]
HIGH
** `ApprovalInbox` GET read routes remain unauthenticated and wildcard-CORS after the GHSA-mxjx-28v… affects [">= 5.12.2, <= 5.13.3"]
MEDIUM
AgentRuntime sandbox path-prefix checks allow file access outside the configured base directory EPSS 0% CVE-2026-58481 affects ["<= 5.12.1"]
MEDIUM
EnvironmentManager.restore() backup ID path traversal copies arbitrary directories into environment… EPSS 0% CVE-2026-58413 affects ["<= 5.12.1"]
MEDIUM
EnvironmentManager.backup() follows symlinked directories and copies files outside the environment … EPSS 0% CVE-2026-58414 affects ["<= 5.12.1"]
MEDIUM
ApprovalInbox HTTP server has no authentication — anyone can approve pending agent actions EPSS 0% CVE-2026-58482 affects [">= 5.0.0, <= 5.12.1"]
tool safety 1 findings · grade factor +5
  1. medium dangerous code

    dynamic exec: eval()

skills & danger signals github-tarball
prompt-surface shipped agent-instruction files + hidden-content / dangerous-code findings — quoted from the analyzed source

analyzed commit 32fb471 · analyzer v33 · 14h ago

skills & prompt files 5

danger signals1

embed badge readme-ready
live risk-grade badge preview [![MCP Observatory risk grade](https://mcpobservatory.com/servers/github:Jovancoding/Network-AI/badge.svg)](https://mcpobservatory.com/servers/github:Jovancoding/Network-AI/security)

Heuristic, inferred signals — false positives (legitimately powerful tools, forks, language ports) are expected. Treat each as "review this", not a verdict. See the ecosystem-wide picture on the security hub, or the fleet security of Jovancoding.