root 2b5bf86b0f fix: bump maxTokens again (800/500) and add conciseness instructions ...

Post-restart logs show model still hitting 600-token ceiling on verbose
responses — truncated JSON at exactly 599-600 output tokens. Average
successful response was ~418 tokens but outliers reach 600+. Added
explicit conciseness instructions to backstory prompt to reduce variance.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

2026-03-10 18:42:56 +00:00

..

src

fix: bump maxTokens again (800/500) and add conciseness instructions

2026-03-10 18:42:56 +00:00

.env.example

feat(llm): add fallbackModel config, remove hardcoded model default

2026-03-09 15:49:48 +00:00

package.json

chore: add better-sqlite3 dependency and saves directory

2026-03-09 14:07:12 +00:00

tsconfig.json

feat: scaffold monorepo with shared, server, and client packages

2026-03-06 18:09:07 -05:00