Files · shell · git · search

Agent tools

The agent does not only talk. It reads, searches, patches, and — when you allow it — runs shell and git.

Eyeban ships a full tool surface inside a native workbench. The cloud agent calls tools; you see the trail and review the patch.

What it can do

  • Read, write, and patch files with reviewable diffs
  • Grep, glob, list, and semantic search over the workspace
  • Shell with a real PTY session, plus Await for long commands
  • Git status, diff, stage, commit, and push when permitted
  • LSP diagnostics and lints
  • Web search and fetch for docs and public pages
  • MCP tools you attach, plus Computer Use on Windows

Guardrails

Ask, Allow safe, and Allow all decide what runs. Plan and Ask modes block mutating tools. Nothing is silent.

What this is not

There is no cloud VM agent, no plugin store, and no “run the whole OS unsupervised” mode. Tools execute on your machine, under your permissions.

The Tools docs page lists the same surface the agent actually has.

Eyeban · shop/checkout.py Balanced
41 def charge(order, wallet):
42     tax = order.subtotal * 0.09
43     total = order.subtotal + tax
44     if wallet.balance < total:
45         raise InsufficientFunds(total)
46     wallet.debit(total)
47     return Receipt(order.id, total)

Split regional tax from the hardcoded rate, and check coverage before debit.

Read 6 files · 2 searches · 1 test

charge still hard-codes 0.09. tax_for(region) exists. I will wire can_cover to the wallet.

12-line patch across two files. Review the diff.

main Agent Balanced You stay in review