Shipeasy
Bugs & Requests

Slack

Run your whole flags & experiments platform from Slack by chatting with @Shipeasy — create gates, configs, kill switches, experiments, alerts and metrics in plain language, then act on everything as it streams back into the channel as interactive cards.

Production readyOn this page · 7 min readUpdated · June 23, 2026Works with · Dashboard · per project

Slack is a full front-end to Shipeasy. Connect it once, then talk to @Shipeasy in plain language to create and change anything — feature flags, dynamic configs, kill switches, A/B experiments, alerts, metrics — and watch it all stream back into the channel as interactive cards you act on without ever opening the dashboard. Bugs, requests and platform events land the same way.

Connect Slack

Add the connector

On the Feedback page open Connectors → New connector, pick Slack, name it, and Continue.

Add to Slack

You're redirected to authorize the Shipeasy Slack app in your workspace. After consent you land back in the configure view.

Pick channels

Choose the feedback channel new bugs/requests post to. Optionally pick a separate notifications channel for platform events — leave it on Same as feedback channel to share one. Tick Enabled and Save.

Meet @Shipeasy — run the whole platform from chat

Mention @Shipeasy in any channel it's in and ask for what you want. There's nothing to memorise — no flags, no syntax. It understands plain language, acts as you (your Slack identity is matched to your Shipeasy email, and you must be a project member), and tells you exactly what it did.

It can do everything the dashboard assistant can — read and change your entire platform:

Ask @Shipeasy to……and it will
Feature flagscreate a flag, roll it out to a %, choose who sees it, turn it on/off
Dynamic configscreate a config, change what it returns
Kill switchescreate one, flip it on or off
A/B experimentsset one up, start or stop it, ship the winning version, read the results
Alertstell you when a number gets too high or too low
Metricsstart tracking something people do
Just asklist what you have, check how a test is doing, "what changed today?"

Create anything, just by asking

@Shipeasy · create anything
@Shipeasy create a feature flag for the new pricing page and roll it to 10%
✅ Created the “new pricing page” flag and rolled it to 10% in prod.
@Shipeasy add a kill switch for payments
✅ Created a “payments” kill switch — it’s off for now.
@Shipeasy set up an A/B test for the new checkout, and measure how often people finish checking out
✅ Drafted the “new checkout” test — an even split between the old and new versions, measuring completed checkouts. Say “start it” to launch.
@Shipeasy let me know when our API error rate gets too high
✅ Created the “high API error rate” alert — I’ll post here the moment it spikes.
No dashboard, no syntax — describe what you want and @Shipeasy creates it, acting as you. Each reply confirms exactly what happened.

It's just as good at reading and changing what already exists:

@Shipeasy · chat
Dana10:41 AM
@Shipeasy which tests are showing a clear winner right now?
ShipeasyAPP10:41 AM
📊 2 of 4 have a clear winner: the new checkout (new version up 12%) and the pricing test (variant B up 6%). The other two need more data.
Ask questions too — @Shipeasy reads your flags, tests, configs and metrics and answers in the thread.
@Shipeasy · chat
Dana10:41 AM
@Shipeasy roll the new pricing page out to half of everyone
ShipeasyAPP10:41 AM
✅ The new pricing page is now live for 50% of users in prod.
One sentence to change a rollout — the change goes live everywhere immediately.
It always acts as you

@Shipeasy runs every action as you — your Slack identity → your Shipeasy email → a project membership check. A non-member gets a quiet "ask an admin to invite you," never silent access.

From chat to channel — what you set up comes back to you

The alert you just asked @Shipeasy to create doesn't disappear into a settings page. When it fires, it posts into your channel as a card — and you control its whole lifecycle right there.

notification · alert
ShipeasyAPP10:42 AM
🔴 API error rate is too high
The API error rate hit 8.7% over the last 24 hours.
View in dashboard
ResolveDismissMute rule
A triggered alert, bordered by severity. Resolve or dismiss the activation, or Mute rule to disable the rule that fired — all write straight back into Shipeasy.

Click Resolve and the same message is rewritten in place — buttons gone, stamped with who acted, the border flipped to green. Dismiss leaves a distinct grey, and Mute rule disables the rule that fired so it stops nagging.

notification · resolved
ShipeasyAPP10:42 AM
🔴 API error rate is too high
The API error rate hit 8.7% over the last 24 hours.
✅ Resolved by @dana
After you click Resolve the same card is rewritten in place — buttons gone, stamped with who acted, border flipped to green. Dismiss would leave a distinct grey instead.

When a specific error caught by see() starts spiking, that lands too — with a little chart of the last hour so you can see how bad it is at a glance, and the same resolve / dismiss / mute controls:

notification · error spike
ShipeasyAPP10:42 AM
🛑 Checkout is failing for users — error spiking
PaymentGatewayTimeout — the payment provider didn’t respond in 8s.
Caught by see() · 47× in the last hour (up from ~3/hr) · 12 users affected.
View in Shipeasy ↗
ResolveDismissMute this error
A specific error caught by see() — its consequence, the raw error, and a sparkline of the last hour. View it in Shipeasy, or resolve / mute it right from the card.

The same loop runs for everything you set up. Start an experiment and, when it reaches significance, the result lands with the actions that matter:

notification · experiment
ShipeasyAPP10:42 AM
📈 The new checkout test has a clear winner
The new version is winning on completed checkouts — up 12.4% vs the old one.
View in dashboard
Ship the new versionStop the testKeep running
An experiment reached significance. Ship the winning group, stop, or keep running — Ship and Stop go through the admin path and check your project membership first.

…and if a guardrail moves the wrong way, you can stop it on the spot:

notification · guardrail
ShipeasyAPP10:42 AM
🛡️ A safety check slipped on the new checkout test
Checkout is 8.1% slower for the new version than the old one.
View in dashboard
Stop the testAcknowledge
A guardrail metric moved the wrong way. Stop the experiment on the spot, or acknowledge to mark it seen.

Flip a kill switch and the channel offers a one-click revert; publish a config to prod and the change is announced with a link:

notification · kill switch
ShipeasyAPP10:42 AM
🔁 The payments kill switch was turned on
Payments are paused in production right now.
View in dashboard
Flip it back off
A kill switch flipped. One click flips it back — gated to project members and routed through the admin path so the change propagates.
notification · config
ShipeasyAPP10:42 AM
⚙️ Pricing settings published to production
Version 7 — published by ana@acme.co.
View in dashboard
Acknowledge
A dynamic config was published to prod. The card links to the change and lets you acknowledge it.

Every action writes straight back into Shipeasy, records an audit entry, stamps who did it, and recolours the card. Here's the full set:

EventCard actions
Alert triggeredResolve · Dismiss · Mute rule (disables the rule that fired)
Error spiking (from see())Resolve · Dismiss · Mute this error (with a trend chart)
Experiment reached significanceShip "<group>" · Stop experiment · Keep running
Guardrail breachedStop experiment · Acknowledge
Kill switch flippedFlip back (one-click revert)
Config published to prodAcknowledge
Agent needs your attentionOpen the issue ↗ · Acknowledge
Weekly digest · New team member(informational — link only)

Choose which of these post to Slack in Settings → Notifications (the Slack column is live once a connector exists).

Privileged actions check your membership

Stopping or shipping an experiment, flipping a kill switch, and muting an alert rule go through the real admin path (so the change propagates) and verify you're a member of the project first. If you're not, only you see a quiet "ask an admin to invite you" — the shared card is untouched for everyone else.

Bugs & requests — the other inbound stream

New bug reports and feature requests post to your feedback channel with the title, reporter, page URL and full details. Each carries a Status ▾ dropdown covering the whole lifecycle.

feedback · open
ShipeasyAPP10:42 AM
🐞 Bug report: Login button does nothing
Reporterdana@acme.coPage/login
Steps to reproduce
Click “Log in” — nothing happens, no dialog.
Open View in Shipeasy ↗
Filed by Shipeasy
A bug report posted to Slack. The blue left border is its status; the Status ▾ dropdown moves it through the lifecycle (a distinct grey for Won't fix), recolouring the border in place.
StatusBorder
Open🔵 blue
In progress🟠 amber
Ready for QA🟠 amber
Resolved🟢 green
Won't fix / Dismiss⚪ grey

Pick one and the card recolours its left border to match — Won't fix is a distinct grey, never the green of Resolved. Link a PR to the item in Shipeasy and the same message is edited in place to show it:

feedback · pr
ShipeasyAPP10:42 AM
🐞 Bug report: Login button does nothing
Reporterdana@acme.coPage/login
Steps to reproduce
Click “Log in” — nothing happens, no dialog.
🔗 Pull request: #128
View in Shipeasy ↗
Filed by Shipeasy
Link a PR to the item in Shipeasy and the same Slack message is edited in place — the PR appears and the border turns amber (work in flight).

…and when it's done, the card stamps who closed it and turns green:

feedback · resolved
ShipeasyAPP10:42 AM
🐞 Bug report: Login button does nothing
Reporterdana@acme.coPage/login
Steps to reproduce
Click “Log in” — nothing happens, no dialog.
View in Shipeasy ↗
✅ Resolved by @dana
Pick Resolved from the dropdown and the card stamps who closed it and turns green — the buttons collapse away.

When an agent gets stuck — open the issue

Unattended agent runs (the ops loop burning down your queue) raise an "Agent needs your attention" card when they're blocked. Open the issue ↗ takes you straight to the item it's stuck on — not a generic dashboard landing — where the full escalation (the summary and the numbered steps the agent left) is shown in context.

notification · agent attention
ShipeasyAPP10:42 AM
🤖 Bug #42 needs your attention
Can’t reproduce the checkout failure without a test payment key.
1. Add a test payment key to the local setup
2. Re-run the checkout test
Open the issue ↗
Acknowledge
When an unattended agent run gets blocked it raises this card. Open the issue ↗ takes you straight to the item it's stuck on, where the full escalation (summary + steps) is shown in context.

Prefer a quick slash command?

For a fast, no-chat shortcut, /shipeasy creates a metric alert from inside a channel — the alert posts back to that channel, and the reply is private to you:

/shipeasy alert <metric> <gt|gte|lt|lte> <threshold> [name]

For anything beyond a quick alert, just talk to @Shipeasy.

Drive Shipeasy from Slack's AI

If you'd rather use Slack's own AI assistant, Shipeasy is also available there as an MCP connection — the same flags, configs, kill switches, experiments and metrics. Your workspace admin enables it once under the Slack App's MCP Server Connection (Slack identity auth); after that, ask Slack's assistant to manage Shipeasy and it calls through to your project under the same membership check.

Channel routing

  • Feedback channel — where new bugs and requests land.
  • Notifications channel — where platform events post (defaults to the feedback channel).
  • Per-alert channel — an alert created in a specific channel posts back to that channel, overriding the project default.

Status & health

The connector in the list shows a status — Needs OAuth, Needs channel, Active, Disabled, or Failing (last dispatch errored). The configure view surfaces the last error and last successful delivery time, so you can tell at a glance whether mirroring is healthy.

The bot needs your email

The @Shipeasy assistant, privileged card actions, the /shipeasy command, and the Slack-AI connection resolve access by your Slack email, which needs the users:read.email scope. If you connected Slack before this was added, click Add to Slack again to re-grant.

Was this page helpful?
✎ Edit this page

On this page