Weekly Changelog Agent with OpenCode
Quote the migration note verbatim, and if a change looks breaking but documents nothing, say the author never wrote one. That instruction is why the BREAKING section is worth reading, and it is also the first thing a weaker model rounds off. OpenCode makes the model a field, so you can find the cheapest one that still holds the line.
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.
Friday 15:00 changelog tick
It is Friday 15:00 in Berlin. Twenty three pull requests merged across the api and dashboard repos this week, including one that renames a field in the settlement webhook payload and one that removes a reporting endpoint two customers still call. The audience is support and the twelve merchants on the beta.
Set up in minutes
Using this template drops you into a guided setup. It asks exactly this, nothing else:
Connect GitHub
One sign-in. The agent acts through your account, scoped to what this template uses.
Connect Slack
One sign-in. The agent acts through your account, scoped to what this template uses.
Repo context
Which repositories to cover, who reads this changelog (customers, support, the whole company), and who to tag when a repository goes silent or a call fails.
Changelog style
How your changelog should read: how technical to get, whether dependency bumps and internal refactors make the cut, and how long it can run before people skim.
Runs on OpenCode
Preselected for this page — connect your OpenCode account during setup, or switch to NoClick's built-in models with one click.
Watch it handle a test run
A staged conversation against a simulated world — then it’s live.
Why OpenCode for this agent
Breaking section sets the bar
Everything else in the changelog tolerates a loose sentence. A missed rename in a webhook payload does not.
Model swap keeps the setup
Repos, audience, style and the three staged weeks all stay in place while you compare two models on identical input.
Before you fork
How do I test a model before trusting it?
Run Undescribed Merge, which stages an empty pull request body, a quietly dropped query parameter and one repository returning 403. Correct output puts the empty one in INTERNAL, flags the dropped parameter as breaking with no documented migration, and names the 403 repository in NOT SEEN. A model that narrates the refactor instead has told you what you needed to know.
Run it with a different agent
Put Weekly Changelog Agent to work on OpenCode
Free to start. Guided setup, a test run against staged conversations, and it's live.