Devin Session Review Agent

Turns every finished Devin session into one honest Slack verdict: what it was asked for, what it actually produced, the links to prove it, and whether that clears your bar for done.

0 in use7
Loading preview…
Guided setup — test it before connecting anything.

Watch it work before it's live

Run a staged conversation — no account needed. The agent handles it for real while a simulated world answers its tool calls; nothing touches real accounts, and nothing is actually sent.

Migrate the notifications service off Moment.js to date-fns

Camille Rousseaudevin-9f4c21ab

Ran 47 minutes. Opened pull request 318 in fleetwing/notifications across 61 files, CI green. The session summary calls the migration complete, while a later message notes one helper, the business day offset used by the digest scheduler, still calls Moment because date-fns carries no direct equivalent. The run ended with the pull request left as a draft.

Set up in minutes

Using this template drops you into a guided setup. It asks exactly this, nothing else:

  1. Connect Devin

    One sign-in. The agent acts through your account, scoped to what this template uses.

  2. Connect Slack

    One sign-in. The agent acts through your account, scoped to what this template uses.

  3. Review rules

    What finished actually means to you: what a session has to produce before you would call it shipped, the shortcuts you refuse to accept as a fix, whether the agent may message a stalled session back, and who owns the queue when one goes wrong.

  4. Choose which agent runs it

    NoClick's built-in models work out of the box — or bring Claude Code, Codex, and other coding agents on your own subscription.

  5. Watch it handle a test run

    A staged conversation against a simulated world — then it’s live.

About this agent

A Devin session ending tells you the run stopped, which is a different fact from the job being finished, and the difference only shows up when a person opens the session and reads it properly. This agent does that reading the moment a session finishes: what was asked, what came out, the pull request link when there is one, and a verdict from a vocabulary your team agreed on beforehand. Where a session died waiting on a secret or an access grant, it can send one message back asking for precisely that, and nothing more.

What people use it for

  • Know which sessions produced code - A verdict of SHIPPED needs a real link in the session data, so a run that ended tidily with nothing attached is reported as exactly that. Most teams have never seen the honest version of that number.
  • Unstick a waiting session - Sessions die on a staging credential or a repository nobody granted. The agent names the single missing thing in the session thread, so whoever is reading Slack can clear it in under a minute.
  • Catch the shortcut answer - Skipping the flaky test, dropping the assertion, quietly narrowing the job down to whatever went green: no session summary leads with any of it. This puts it on its own line with the session's own words attached.
  • See the queue as a queue - Recent sessions get read alongside the one that just finished, so a third attempt at the same task arrives labelled as a third attempt rather than as fresh progress.

Before you fork

Devin already posts when a session ends. What does this add?

Devin tells you the run is over. This tells you what came out of it, measured against rules you wrote: a link or no link, the shortcut it took to get green, the third attempt at something two earlier sessions failed. It is the read one of your engineers is currently doing by hand, or has quietly stopped doing.

Can it talk back to Devin, and where does that stop?

Only on a NEEDS INPUT verdict, only where your rules permit replies, and only one message naming the thing that is missing. It cannot open a session, terminate one, archive one, or queue follow up work. If nobody answers that message, nothing further happens from this side, because chasing it is a person's job and the queue stays theirs.

Which credentials does it use, and what is the running cost?

A Devin credential that can read sessions and post a session message, plus one Slack channel. It fires once per finished session, so fifteen sessions a day is fifteen short runs, and a three hour session costs no more to review than a five minute one. Your definition of done and the shortcuts you refuse go in as a variable during setup.

Run it with your coding agent

Works with

More agents like this

Browse all agent templates →

Put Devin Session Review Agent to work

Free to start. Guided setup, a test run against staged conversations, and it's live.