https://mcp.enokilabs.ai/mcp. What each does and the scope it needs:
Discover what we test
Configure a target
Trigger & poll assessments
Read findings & remediate
list_findings pagination
list_findings returns one page at a time, newest first: 200 findings by
default, 200 at most, so a large workspace is read in pages rather than in
one response that floods the calling agent’s context.
Every page carries:
To read a workspace to completion, call
list_findings, and while the result
carries next_cursor, call again with cursor set to it and the same
filters. The order is stable under concurrent ingest: a finding that lands
mid-walk appears above every cursor already handed out, so no finding is
repeated or skipped. A cursor is bound to the filters it was minted under, so
passing it back with different ones is rejected as an invalid argument rather
than resuming part-way through a list it never belonged to; a malformed or
edited cursor is rejected the same way. limit may change between pages.
total_matched is counted after every filter, min_severity included, so the
remainder it implies is the remainder you will get. truncated is never derived
from it: the flag says whether this page is the last, and a page is never
reported complete when it is not.
Full parameters & responses
Full inputs and response shapes travel with each tool overtools/list (the
create-target config schema is the create_target_schema resource). For a guided
walkthrough, call get_setup_guide or see Connect your agent.