ThumbGateThumbGate

What AI coding agents tried to do in November 2026

This is data from our own fleet, not a survey and not customer data. It covers 2026-11-01 to 2026-11-17 on one Mac where ThumbGate gates our own coding agents. It is a small sample, and we say so below.

34risky actions stopped or flagged (30 denied, 4 warned)
124agent actions the gate checked (98 allowed, 24 denied, 2 warned)
1,995agent sessions registered by our agents

What the gate stopped or flagged, by action type

Action typeDeniedWarnedTotal
Commit on a PR branch before review threads were verified14014
Dependency added to package.json011
Edit outside the declared task scope303
Force-push606
High predicted workflow risk before execution022
Pull request opened without permission evidence101
Repeat of a high-risk action ThumbGate had already been told was bad505
Secret material in a tool call101
Unauthorized network egress011

An action type is the rule that fired. Denied means the tool call did not run. Warned means it ran after a warning.

Dangerous-looking shell commands, and what the gate did

Of 80 shell commands the gate engine evaluated, these counts are commands whose text contains each pattern. Zero rows are real zeros in this window.

Command patternSeenAllowedWarnedDenied
rm -rf (recursive force delete)0000
git push --force5005
git reset --hard0000
Reading a secrets file or keychain entry0000
DROP TABLE / DROP DATABASE0000
curl or wget piped into a shell0000
npm publish0000
git --no-verify0000

A command can match a pattern and still be allowed, for example rm -rf on a build folder inside the declared task scope. Pattern matching counts text, so a command that only mentions a pattern (an echo, a commit message) is counted too. The rule table above counts rules that fired, so it can differ from this table when a rule matches a form of the command our text pattern does not.

What the gate allowed

Of 124 agent actions (shell commands, edits, file reads and writes) the gate engine evaluated, 98 were allowed, 24 were denied and 2 produced a warning. Separately, agents made 537 calls to ThumbGate's own tools (feedback, recall, gate checks); 532 were allowed and 5 were denied.

Agent action toolChecked
Bash78
Edit33
cronjob_manage3
Write3
patch2
Read2
terminal2
cronjob_manage.update1

How we counted, and what this does not show

Run it on your own agents

npx thumbgate init --agent claude installs the gates; see pricing for what Pro adds.