Summary

This adds a Context Bloat Analyzer finding to codeburn optimize so users can spot sessions where the effective input/cache footprint is much larger than the useful output. The goal is to make it obvious when a fresh thread or tighter context handoff would likely be cheaper than carrying everything forward.

• adds a new detectContextBloat() detector in the optimize pipeline
• measures effective input/cache tokens using the existing cache pricing assumptions: cache reads are discounted, cache writes use the cache-write multiplier
• flags sessions only when effective input/cache tokens are both large and disproportionate to output
• shows the top context-heavy sessions, their input/output ratio, and sharp growth from the previous project session
• caps extreme ratio display as 1000+:1 so zero-output sessions do not produce noisy numbers
• keeps candidate ordering deterministic for stable previews and tests
• documents the detector in README and CHANGELOG

Detection model

The detector is intentionally conservative:

• effective input/cache floor: 75,000 tokens
• minimum effective input/output ratio: 25:1
• token-savings estimate: context above a healthier 15:1 ratio
• high impact when there are at least 3 candidates or at least 500K effective input/cache tokens across candidates
• growth is measured against the immediately previous session in the same project, even if that previous session is below the reporting threshold

This keeps raw cache-read-heavy sessions from being overstated while still catching sessions that are expensive in practice.

Validation

• npx vitest run tests/optimize.test.ts
• npm run build
• npx vitest run
• node dist/cli.js optimize --help
• git diff --check

npx tsc --noEmit still fails on the existing Copilot provider type errors in src/providers/copilot.ts, outside this diff. The same failure is present on origin/main.

thats all i got for the moment :)