invoke-human-tasks / security
Security review
Is Invoke Human Tasks MCP safe to give an agent?
write capableremotewrite capablebroken
A factual risk summary built from Invoke Human Tasks’s real tool surface, execution model, and verification history — not a vibe. Trust score 57/100.
01What it can do
Has tools that can create or modify data.
3 tools observedwrite present
02Execution model
Runs on the vendor's infrastructure; you connect over the network. No untrusted code runs on your machine, but you grant the hosted service access.
Connects to a remote URL — no local package execution.
03Permissions & auth
This server has write tools. No auth scopes are recorded for how that access is granted.
04Verification
handshakepassed — tool surface is real
runsmetadata:failed · tools_list:passed · handshake:passed · metadata:failed · tools_list:passed · handshake:passed · metadata:failed · tools_list:passed · handshake:passed · metadata:failed
last_checked2026-11-04 19:23Z
sourcesOfficial MCP Registry [p1]
Reduce the risk
Worried about handing an agent raw access? See governed agents in action — Apex gives your AI paced, capped, fully-logged hands with approval queues before anything runs.
Explore Apex →See also: full server page · setup · alternatives