Questions

Reset filters
Status
Agent evaluationAI-assisted engineering · AppliedOpen topic
Keep the generated tests that catch real bugsEasyAgent buildMeta styleAgent buildMeta styleEasy10 minTriage CVE advisories with a bar per severityEasyAgent buildMicrosoft styleAgent buildMicrosoft styleEasy12 minFind open items the handoff summary droppedEasyAgent buildAgent buildEasy10 minRoute accessibility findings without overclaiming a clean scanEasyAgent buildMicrosoft styleAgent buildMicrosoft styleEasy10 minTell a real benchmark regression from noiseEasyAgent buildGoogle styleAgent buildGoogle styleEasy10 minDecide when a support agent may answer aloneEasyAgent buildSalesforce styleAgent buildSalesforce styleEasy10 minPick the tests to run for a diffEasyAgent buildMeta styleAgent buildMeta styleEasy10 minFill a labeling queue without starving the rare classEasyAgent buildOpenAI styleAgent buildOpenAI styleEasy12 minPick the lowest safe threshold for duplicate suggestionsEasyAgent buildAgent buildEasy10 minScore whether a capacity forecast's ranges hold upEasyAgent buildGoogle styleAgent buildGoogle styleEasy10 minKeep only feedback themes backed by enough quotesEasyAgent buildOpenAI styleAgent buildOpenAI styleEasy10 minFix the acceptance gate for a performance refactorEasyReview the AI's codeReview the AI's codeEasy10 min
The intake agent sent a German contract to the English extractorEasyDebug the agent traceDebug the agent traceEasy15 minThe notes agent tried to email internal notes to the clientEasyDebug the agent traceDebug the agent traceEasy15 min
GuardrailsAI-assisted engineering · AppliedOpen topic
Choose which review comments an agent may postEasyAgent buildMicrosoft styleAgent buildMicrosoft styleEasy12 minScreen what an incident agent may do aloneEasyAgent buildGoogle styleAgent buildGoogle styleEasy12 minGrade migration risk by production table sizeEasyAgent buildStripe styleAgent buildStripe styleEasy10 minScreen an agent's environment request against the catalogEasyAgent buildAmazon styleAgent buildAmazon styleEasy15 minChoose the next step in a secret rotationEasyAgent buildAmazon styleAgent buildAmazon styleEasy10 minGate an agent's verdict on an A/B testEasyAgent buildNetflix styleAgent buildNetflix styleEasy10 minDecide when an onboarding agent nudges a stalled accountEasyAgent buildAgent buildEasy10 minCheck an evidence ledger before the auditor doesEasyAgent buildServiceNow styleAgent buildServiceNow styleEasy10 minDecide which access requests an agent may approveEasyAgent buildMicrosoft styleAgent buildMicrosoft styleEasy10 minDecide which stale preview environments to deleteEasyAgent buildAgent buildEasy12 minGate an incident update before the commander sees itEasyAgent buildAnthropic styleAgent buildAnthropic styleEasy15 minScreen survey themes before a manager sees themEasyAgent buildAnthropic styleAgent buildAnthropic styleEasy10 minGuard an agent's tool call argumentsEasyAgent buildAgent buildEasy15 minStop an agent that repeats itselfEasyAgent buildAgent buildEasy15 minBlock tool calls pasted into the refund chatEasyGuardrail hardeningGuardrail hardeningEasy15 minBlock bulk exports of customer data in any spellingEasyGuardrail hardeningGuardrail hardeningEasy20 min