Aggrete is an MCP proxy that sits between AI assistants and your connectors. It checks each tool call against a YAML policy and per-user memory, and refuses the request that would leak a secret, cross an information barrier, over-share health data, complete a forbidden combination, or write out after reading untrusted content, before the upstream is ever contacted.
pip install aggrete && aggrete scan
pip install aggrete && aggrete scan--fix writes the policyuvx aggrete --demodocker run --rm ghcr.io/aggrete/aggrete --demoOne command reads every MCP client config on a machine, connects to each server, and prints what no per-server scanner does: tool combinations across servers that form a leak path, hidden instructions in descriptions, secrets in configs, servers that bypass the proxy. --fix writes the config and policy that close them.
Each decision names the assistant or task agent that made the call and the person it acted for. Memory still follows the person, so using two agents is not a way around a rule. The console shows the roster and the agents-per-person ratio.
Why it mattersFourteen widely discussed incidents, each marked stopped, partial or irrelevant, with the checks that demonstrate it. Two stopped, three partial, nine where a proxy is the wrong control and we say so.
See the verdictsEverything it does, in plain words, with what engineers call it in grey. Hover any line for a sentence more.
Every security tool says it blocks these. Aggrete ships the test. Sixteen checks drive the real engine and map onto the lists buyers ask about, and a black-box mode scores any MCP gateway from the outside.
pip install aggrete && aggrete conformanceof 10 controls, 0 failing
of 10 controls, 0 failing
of 14 controls, 0 failing
of 18 controls, 0 failing
Two real runs over MCP. The right-hand column is the same fixture connector with no policy layer in front of it. Full report, every control, and how to point it at your own gateway →
The pre-call gate, the accumulator, the audit chain, and the rule types.
connectorsSlack, GitHub, Drive, Jira, Salesforce and Notion, governed the same way.
handbook to policyPaste your handbook and watch clauses turn into enforceable YAML rules.
get startedInstall, write the policy, wire the upstreams, and pick how to run it.