feat: Phase 1 improvements - Cost tracking, Eval system, Context management - #8

Merged

SevenBT merged 11 commits into

mainfrom

phase1-critical-fixes

Sep 10, 2026

Commits on Sep 6, 2026

fix: resolve 4 critical bugs preventing demo

SevenBT
and
claude
committed
bf49c52
docs: correct factual claims in README and design docs

SevenBT
and
claude
committed
d8db1b9

Commits on Sep 7, 2026

feat: add token cost calculation and tracking

SevenBT
and
claude
committed
7357077

Commits on Sep 8, 2026

feat(eval): connect LLM-as-Judge and create 30-case test suite

SevenBT
and
claude
committed
020bcf2
feat: replace token heuristic with tiktoken for precise context management

SevenBT
and
claude
committed
640669a

Commits on Sep 9, 2026

docs(eval): add W2.1 evaluation report with 31 cases

SevenBT
and
claude
committed
1859a4b

Commits on Sep 10, 2026

fix: correct QConfig.get() usage - no default parameter allowed

SevenBT
and
claude
committed
89da079
feat: add user-configurable model pricing (JSON file)

SevenBT
and
claude
committed
587f6c4
docs: remove reference to moved llm-evaluation.md file

SevenBT
and
claude
committed
5c6c1e5
test: fix imports after consolidator refactoring

SevenBT
and
claude
committed
820c4d0
fix: restore approvals parameter removed in bug fix

SevenBT
and
claude
committed
114b591