Blocking the lethal trifecta: the one attack only the MCP layer can stop
An agent with read-access to secrets, an exfiltration channel, and untrusted input is
the canonical agentic-AI attack. Each tool call is individually authorized — the danger is the sequence,
read_credentials → write_external, and eunox’s sequenceBlock condition is the only practical
place to break that chain.
Read →
From Observation to Policy: Drafting a Manifest From What Your Agent Actually Did
The proxy is the easy part — the blank page where you write the first policy is the
real adoption cliff. eunox suggest reads the wiretap tape and drafts a manifest grounded in
observed usage, with allowedValues conditions built from the argument values the agent actually used.
Read →
One Proxy, Many Upstreams: Governing Every MCP Server From a Single Gateway
Running one policy proxy per MCP server doesn’t scale past a couple of servers. The
gateway fronts every upstream from one process — each on its own route, with its own versioned manifest, all
writing to one signed audit tape.
Read →
The prompt injection problem: why every AI agent needs a policy layer
Why fixing prompt injection inside the model doesn’t work, and why the only reliable
defense is a policy layer at the structured tool call.
Read →