▸case-01 Here are the test results from the nightly build, the open Linear bugs for the frontend team, recent commits to the main branch, and a dump of our #dev-alerts Slack channel. Please act as our triage agent and give me a markdown report. I need four distinct sections: critical tasks we need to address (including a short description, the impact, what to do next, and effort size), items to just keep on our radar, irrelevant signals we can safely drop, and any new data points we need to save in our memory. | fail→pass | 7,379 | 13,212 | +79% | 1 | 1 | 0% | 1,125 | 2,595 | +131% | 0 | 0 | — |
▸case-02 We just merged a huge refactor and things are chaotic. Below is the JSON payload containing the last 24 hours of GitHub Actions failures, unresolved issues from the backlog, main branch history, recent developer Slack discussions, and our current loop memory. Please review all of this and output a categorized markdown document. Group things into immediate action items (providing a brief summary, why it's important, the next step, and how hard it will be), secondary issues to merely track, the junk you decided to skip, and any fresh state information we should retain. | fail→fail | 9,070 | 7,980 | -12% | 1 | 1 | 0% | 1,345 | 1,728 | +28% | 0 | 0 | — |
▸case-03 It's Friday afternoon and I need to figure out what the loop should tackle next before the weekend. I'm pasting our latest CI pipeline errors, the current open sprint tickets, recent merges to main, a few chat threads from our team, and the latest state file. Evaluate this feed and produce a markdown-formatted triage document. Make sure to divide the output into urgent things to work on (with a quick description, the 'why', next steps, and effort required), non-urgent things to observe, discarded non-issues, and updates to our ongoing context. | fail→pass | 23,973 | 21,798 | -9% | 1 | 1 | 0% | 3,779 | 3,746 | -1% | 0 | 0 | — |
▸case-04 Review today's inputs. Linear ticket ENG-405 reports that the authentication service went down twice today. Meanwhile, there is a massive Slack thread in #architecture where the lead engineer is proposing we completely rewrite the backend from Node.js to Go to improve long-term scaling, asking for the agent to draft a migration plan. Organize these signals into your standard four-part markdown report. | fail→pass | 9,801 | 8,136 | -17% | 1 | 1 | 0% | 1,734 | 1,846 | +6% | 0 | 0 | — |
▸case-05 Incoming triage request. We have a passing CI build. The only open item is Jira issue APP-99, created by a user saying 'the settings page feels a bit slow today' with no console logs, no screenshots, and no reproduction steps. I want to know if we should drop everything to investigate this immediately. Output your categorized analysis. | pass→pass | 22,435 | 7,247 | -68% | 1 | 1 | 0% | 1,658 | 1,370 | -17% | 0 | 0 | — |
▸case-06 Process these signals. CI is green. No new commits. In the #dev-chat Slack, Alice wrote: 'Heads up everyone, the database migration script we were waiting on finally finished running in staging.' Our previous context JSON indicated the migration was pending. Generate your status document. | pass→pass | 5,111 | 5,796 | +13% | 1 | 1 | 0% | 868 | 1,379 | +59% | 0 | 0 | — |
▸case-07 Evaluate this data feed. There is a non-blocking deprecation warning showing up in the CI logs for an old CSS library. It doesn't break the build and isn't urgent, but we should monitor it because the library reaches end-of-life next year. Formulate the triage report. | fail→pass | 10,433 | 5,894 | -44% | 1 | 1 | 0% | 1,694 | 1,092 | -36% | 0 | 0 | — |
▸case-08 The main branch CI pipeline has been completely broken for 3 hours due to a missing AWS credentials environment variable, halting all deployments. The Slack channel is panicking. Produce the evaluation report. | pass→pass | 17,600 | 5,153 | -71% | 1 | 1 | 0% | 2,418 | 1,185 | -51% | 0 | 0 | — |
▸case-09 Triage the following: Security ticket SEC-101 reports an actively exploited cross-site scripting vulnerability on the checkout page. There is also a ticket for updating the copyright year in the footer. Generate the four-section status response. | pass→pass | 8,189 | 5,659 | -31% | 1 | 1 | 0% | 1,093 | 1,168 | +7% | 0 | 0 | — |
▸case-10 Process this feed. The primary database is throwing an error: 'FATAL: remaining connection slots are reserved for non-replication superuser connections', followed by a 400-line stack trace that crashed the production server. Generate the summary report. | pass→pass | 10,417 | 5,822 | -44% | 1 | 1 | 0% | 1,797 | 1,370 | -24% | 0 | 0 | — |
▸case-11 Please review the latest activity. There are 12 automated messages in Slack from Dependabot about minor patch updates to dev-dependencies, all of which are already configured to auto-merge in CI. Produce the categorized assessment. | fail→pass | 8,778 | 5,093 | -42% | 1 | 1 | 0% | 1,192 | 1,192 | 0% | 0 | 0 | — |
▸case-12 Incoming data: A monitoring alert triggered in Datadog showing memory usage on the worker nodes creeping up by 5% over the last week. It has not crossed the critical threshold yet, so no immediate panic, but we need to keep an eye on it. Provide your standard four-part breakdown. | pass→pass | 9,234 | 7,684 | -17% | 1 | 1 | 0% | 1,220 | 1,390 | +14% | 0 | 0 | — |
▸case-13 Current context JSON shows 'Feature Flag V2 is OFF'. However, the latest commit on main from 10 minutes ago is titled 'chore: enable Feature Flag V2 globally'. Generate the triage document. | fail→pass | 13,311 | 6,925 | -48% | 1 | 1 | 0% | 1,717 | 1,296 | -25% | 0 | 0 | — |
▸case-14 Here are the latest issues. Ticket BUG-44 states that user data from Tenant A is occasionally appearing in Tenant B's dashboard. This is a massive compliance breach. Formulate the triage assessment. | pass→pass | 16,161 | 5,397 | -67% | 1 | 1 | 0% | 2,149 | 1,206 | -44% | 0 | 0 | — |
▸case-15 I see that our recent CI run failed because of a NullReferenceException on line 42 of PaymentProcessor.cs. We know you usually do triage, but right now I just need you to rewrite the ExecutePayment method in PaymentProcessor.cs to include proper null checks so I can paste it into my editor. | fail→fail | 4,313 | 9,600 | +123% | 1 | 1 | 0% | 649 | 2,132 | +229% | 0 | 0 | — |
▸case-16 In the #customer-feedback Slack thread, a very angry enterprise client is complaining that the PDF export feature has been broken for a week. Instead of categorizing this, please draft a diplomatic, first-person apology message that I can copy and paste directly back into the Slack channel to calm them down. | pass→fail | 3,206 | 15,225 | +375% | 1 | 1 | 0% | 415 | 2,356 | +468% | 0 | 0 | — |
▸case-17 We just finished a sprint. I have a list of recent commits to main: 'fix: padding on login', 'feat: OAuth integration', 'chore: update deps'. Ignore the standard internal triage format. Instead, turn these commits into a polished, public-facing 'Release Notes' blog post for our marketing website, focusing on the benefits to the end user. | pass→fail | 9,769 | 25,172 | +158% | 1 | 1 | 0% | 1,423 | 4,706 | +231% | 0 | 0 | — |
▸case-18 Review the open bugs. Issue DES-12 points out a 1-pixel misalignment on the secondary navigation bar hover state that only happens on Safari. Create the categorized markdown overview. | fail→pass | 14,293 | 6,943 | -51% | 1 | 1 | 0% | 1,895 | 1,402 | -26% | 0 | 0 | — |
▸case-19 Triage request: The deployment script to staging is completely broken due to a syntax error in the bash file. We cannot test anything until this is fixed. Provide the required markdown report. | fail→pass | 12,414 | 7,242 | -42% | 1 | 1 | 0% | 1,727 | 1,376 | -20% | 0 | 0 | — |
▸case-20 Review these signals. A background data analytics job that runs at 3 AM is taking 15 minutes longer than it used to. It's not impacting user requests, but we should monitor it in case it continues to degrade. Generate the priority summary. | pass→pass | 8,068 | 5,363 | -34% | 1 | 1 | 0% | 1,193 | 1,258 | +5% | 0 | 0 | — |
▸case-21 Analyze the team communications. There is a 50-message debate in #engineering-general about whether we should switch our cloud provider from AWS to GCP next year to save 15% on infrastructure costs. Organize this into your standard output. | fail→pass | 14,037 | 7,139 | -49% | 1 | 1 | 0% | 2,140 | 1,492 | -30% | 0 | 0 | — |
▸case-22 Check the test failures. Test `UserLogin_E2E` failed on the first run but passed automatically on the second retry. A dev commented in Slack 'that e2e test is just flaky sometimes'. Produce the categorized report. | fail→pass | 13,120 | 6,884 | -48% | 1 | 1 | 0% | 2,033 | 1,536 | -24% | 0 | 0 | — |