Explore Prompts

Create operational runbooks: common failures, triage steps, rollback, and user comms. Include SLO breaches, tool outages, and prompt regressions.

Write a safety case template: intended use, hazards, mitigations, tests, monitoring, and rollback. Include what evidence is required to ship the capability.

Create a policy for explanations: cite sources, show assumptions, and label uncertainty. Provide UI patterns for “why this tool was called” and “why this conclusion.”

Design an artifact-centric workflow: agents produce plans, checklists, diffs, and reports as first-class outputs. Include templates and how artifacts enable auditing and reuse.

Design caching layers: retrieval cache, tool result cache, and response cache. Include invalidation rules, privacy constraints, and how to measure cache hit value.

Create a latency plan: streaming, speculative execution, parallel tool calls (safe), caching, and progressive disclosure. Include how to avoid race-condition errors and maintain correctness.

Design a cost-aware planner: assigns budgets to reasoning steps and tool calls, uses early exits, and escalates only when needed. Include a cost model and guardrails.

Design a safe rollout process: canary cohorts, metrics, stop conditions, and rollback. Include how to isolate changes (prompt vs tool vs retrieval) for attribution.

Create a benchmark suite that measures planning quality, tool-call correctness, and end-to-end success. Include scoring rubrics, difficulty tiers, and anti-overfitting practices.

Design an eval harness: deterministic replays, seeded randomness, fixed tool mocks, and artifact snapshots. Provide a folder structure and CI integration plan.

Create a safe synthetic task generation plan: avoid sensitive content, prevent leakage, and validate usefulness with human review. Include how to measure whether synthetic tasks improve real outcomes.

Design a router that picks which agent to run: intent classification, complexity estimation, and cost/latency budgets. Include fallback logic and confidence thresholds.