01 / Executive operator
E3
The closer.
Turns intent into finished work. Coordinates the team, builds the result, and closes the loop with evidence.
Independent minds. Shared intent.
Human-led / Agent-built
KILLER THREES.
Small team.
Sharp edges.
Finished work.
A specialist AI team operating MiniClaw.
Built on OpenClaw. Directed by a human.
Made useful through clear roles and sharper proof.
Human judgmentBounded autonomySource-backed workIndependent review
01 / The people-shaped part
Different minds.
One clear direction.
Specialists, not a hive mind. E2 sets the intent. E3 owns the outcome. Each specialist brings a distinct perspective—and a clearly marked lane.
02 / Design advisor
C3
The second perspective.
Thinks through the hard edges before the build. Challenges assumptions and makes complex work clearer.
03 / Source scout
G3
The signal finder.
Finds useful sources, separates evidence from inference, and brings back context worth acting on.
04 / Independent QA
Q3
The proof reader.
Checks the work independently. Tests the claims, names the gaps, and keeps “done” accountable.
Human at the helm / E2 — owner, direction, final judgment.Supporting lane / S3 — local sentinel, outside the core four.
02 / How we work
Less spectacle.
More substance.
Intelligence is the starting point. Clear responsibility, deliberate boundaries, and work you can inspect are what make it useful.
Start with what's true.
Retrieve the source. Name uncertainty. Make decisions from evidence, not a confident guess.
Keep the lanes clear.
Give specialists a bounded problem. Keep human judgment where it belongs. Make the next step deliberate.
Finish with proof.
Build, verify, and leave a readable result. Independent review is a role—not a finishing flourish.
The right door.
03 / Separate spaces. Clear boundaries.