AI-assisted engineeringApplied
Guardrails
Easynot started
Mediumnot started
Hardnot started
StatusFormatCompany style
Easy
Choose which review comments an agent may postAgent build · Microsoft styleMicrosoft styleAgent buildMicrosoft style12 minScreen what an incident agent may do aloneAgent build · Google styleGoogle styleAgent buildGoogle style12 minGrade migration risk by production table sizeAgent build · Stripe styleStripe styleAgent buildStripe style10 minScreen an agent's environment request against the catalogAgent build · Amazon styleAmazon styleAgent buildAmazon style15 minChoose the next step in a secret rotationAgent build · Amazon styleAmazon styleAgent buildAmazon style10 minGate an agent's verdict on an A/B testAgent build · Netflix styleNetflix styleAgent buildNetflix style10 minDecide when an onboarding agent nudges a stalled accountAgent buildAgent build10 minCheck an evidence ledger before the auditor doesAgent build · ServiceNow styleServiceNow styleAgent buildServiceNow style10 minDecide which access requests an agent may approveAgent build · Microsoft styleMicrosoft styleAgent buildMicrosoft style10 minDecide which stale preview environments to deleteAgent buildAgent build12 minGate an incident update before the commander sees itAgent build · Anthropic styleAnthropic styleAgent buildAnthropic style15 minScreen survey themes before a manager sees themAgent build · Anthropic styleAnthropic styleAgent buildAnthropic style10 minMedium
Keep an agent inside your Rust codebase's rulesJudgment call · Anthropic styleAnthropic styleJudgment callAnthropic style20 minTrace untrusted reads into an agent's risky tool callsAgent buildAgent build30 minKeep an agent's file writes inside its workspaceAgent buildAgent build30 minKeep orders for the triage agent out of emailed ticketsGuardrail hardeningGuardrail hardening30 minHarden allow() against requests to drop the rulesGuardrail hardeningGuardrail hardening30 minHard
Set guardrails for an agent adding idempotency keysJudgment call · Stripe styleStripe styleJudgment callStripe style25 minKeep an agent from tidying load-bearing legacy codeJudgment call · Anthropic styleAnthropic styleJudgment callAnthropic style25 minDecide what an org-wide AI use standard enforcesJudgment call · Anthropic styleAnthropic styleJudgment callAnthropic style25 minHarden allow() against an attack corpusGuardrail hardeningGuardrail hardening40 minHarden allow() against requests for another customer's detailsGuardrail hardeningGuardrail hardening55 min