Prompts by industry
Knowledge Work landing page prompts
2 prompts tagged Knowledge Work. Each one was run against Claude 4.6 Sonnet, Codex and GLM-4.6 before it shipped — the fidelity score you see is the average of the recorded runs, with screenshots and failure notes on the prompt page. 2 are verified; the rest are beta and say so.
2 prompts2 verifiedavg fidelity 89/1006 recorded runs
Open prompt →
AI answer panel with citation chips
trustworthy, sourced, clean
VerifiedReactHTML
90/100 · 3 runsOpen prompt →
AI answer feedback and retry interface
direct, forgiving, human
VerifiedReact
87/100 · 3 runsThis page is derived from the prompt catalog — nothing here is written by hand, so the counts can never drift from the data they describe.