The Bug Class AI Coding Agents Keep Introducing (and How We Started Catching It in CI)

Momcilo Savic — Wed, 26 Aug 2026 20:57:02 +0000

The pattern

AI coding agents are good at producing a diff that works in the
narrowest sense — the function still returns what the test expects. What
they're not reliably good at is preserving properties nobody wrote a test
for in the first place.

The two we kept running into: an authorization check quietly dropped
during an agent-driven refactor (nothing failed, because no test covered
who was allowed to call the route — only that the route worked), and a
rewritten query that behaved fine against a small dev dataset and
full-table-scanned the moment it hit production data. Neither shows up in
CI as it exists today. Both show up in code review only if the reviewer
happens to look at exactly the right five lines out of a few hundred.

What we built

Agent Code Merge Gate
is a free GitHub Action, now live on the GitHub Marketplace,
that runs on every pull request and scans the diff specifically for those
two regression classes. It runs an offline heuristic pass (fast, no
external call) plus one AI-backed pass for a short Executive Summary, and
posts a single comment back to the PR that updates on every push rather
than piling up duplicates.

Deliberately narrow scope — it's not trying to be a general linter. It
covers the two failure modes we found ourselves manually re-checking for
once AI-generated PRs became the majority of our merge volume.

Wiring it into CI

Three lines in a workflow file:

- name: Agent Code Merge Gate
  uses: avalonlabs-platform/agent-code-merge-gate@v1.0.0
```
{% endraw %}


No signup and no config needed for the default behavior. Two inputs worth
knowing about: {% raw %}`fail-on-critical: true` turns a CRITICAL finding into an
actual failed check instead of just a comment, and `comment-on-pr: false`
if you'd rather build your own notification from the raw `status` output.

## What's next

Right now it's diff-scoped — it sees what changed in this PR, not the
whole repo's history of how that code got there, which limits how much
context it can reason about. Whole-repo context is the obvious next step,
and it's also where this stops being a free CI script and starts being a
product decision — worth its own post once it's built rather than
speculated about here.

If you've hit a different bug pattern that seems to show up
disproportionately in AI-generated PRs, I'd genuinely like to hear about
it — that's exactly the kind of thing worth building detection for next.

DEV Community: Momcilo Savic

The Bug Class AI Coding Agents Keep Introducing (and How We Started Catching It in CI)

The pattern

What we built

Wiring it into CI