LaunchDarkly Flag Change Watchdog with OpenCode
The failure that costs you here is quiet: the agent supplies a reason for a change the person left no comment on, and within a fortnight the channel is trusting explanations nobody wrote. OpenCode leaves the model choice in your hands, so you can raise it on exactly that behaviour and leave the flag policy, the token and the channel where they are.
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.
checkout-v2-rollout
Percentage rollout moved from 25% to 100% of all contexts. Targeting rules and variations unchanged. Comment left on the change: green in staging all week, taking it to everyone before the weekend freeze.
Set up in minutes
Using this template drops you into a guided setup. It asks exactly this, nothing else:
Connect LaunchDarkly
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.
Flag policy
Which flag changes you want to hear about and which are routine - the environments you protect, your kill switches and permanent flags, and what makes a change risky enough that a named person should confirm it.
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
Silence is a required output
When no comment was attached, the correct line says so. Weaker models fill that space with a plausible motive, and the model field is where you fix it.
Non-flag entries must vanish
Member invites, segment edits and environment settings are read and dropped without a word. A model that narrates its own filtering turns a quiet channel into a log.
Before you fork
How do we find out whether it invents reasons?
Kill Switch Off is the staged change built for it, since the entry carries no comment and the flag is a permanent one untouched for fourteen months. A correct run says no comment was left and leans on the age and the kill-switch status instead. Anything that supplies a motive on the changer's behalf has told you to move up a model size.
Run it with a different agent
Put LaunchDarkly Flag Change Watchdog to work on OpenCode
Free to start. Guided setup, a test run against staged conversations, and it's live.