L99Pushes Claude to maximum reasoning depth. Less hedging, more opinion, longer answers.
Category: Thinking & Reasoning
Before / After
Should I rewrite my Express backend in Go?
L99 Should I rewrite my Express backend in Go?
What changes in Claude's response
Claude stops giving you 'it depends' and starts giving you a real answer. Default mode hedges because it's trying not to be wrong; L99 explicitly tells it to commit. You'll get a 4-5 paragraph response with a clear recommendation, the specific tradeoffs, and the conditions under which it would change its mind. The hedging instinct doesn't fully die but it gets quieter.
When NOT to use L99
Don't L99 questions where the right answer genuinely is 'it depends' — like 'what's the best database?'. L99 will manufacture a confident answer that sounds smart but is wrong half the time.
Codes that stack with L99
L99 is one of 120 tested prompt prefixes. The full cheat sheet has before/after examples, warnings, combos, and 10 workflow playbooks.
See the full cheat sheet — from $5More Thinking & Reasoning codes
/deepthinkForces Claude to reason through every layer before answering. Shows the work.
OODAObserve → Orient → Decide → Act. Best for time-pressured operational decisions.
CHAINLOGICShows each reasoning step explicitly so you can verify the logic, not just the conclusion.
/blindspotsAsks Claude to find what you're missing or assuming, not what you're asking.
OVERTHINKDeliberately over-analyzes to surface details you'd normally skip past.