Releases: waswrsis/Quant-Strategy-Tokenizer
Releases · waswrsis/Quant-Strategy-Tokenizer
v0.4.0 Agent Handoff Final
Final archival release for QST as an agent-ready research prototype.
This release preserves QST as a typed, canonical strategy-record system for research, review, coverage measurement, and agent handoff.
Highlights:
- Typed GKR records on
qst-ir/0.4andqst-canonical/0.4 - Deterministic validate / hash / canonicalize workflows
- Token surface governance and conformance coverage
- Coverage Frontier v0.3 matrix, report, dogfood set, and publication gate
- Active Stage 3C agent prompt pack and final handoff documentation
- Qlib partial workflow adapter proof via
qst adapter qlib import
Qlib adapter boundary:
- Imports selected Qlib workflow YAML into candidate QST record-layer GKR and coverage JSON
- Does not import Qlib
- Does not run qrun
- Does not train models or run inference
- Does not execute backtests
- Does not connect to brokers or exchanges
- Does not claim lossless Qlib conversion
Not included:
- live trading
- broker or exchange integration
- production execution
- full backtesting
- Qlib runtime replacement
- profitability claims
Primary tag commit: 082e3b9a6c61934eb8cc106d4fd4bddccd80bbc2.
Note: main also includes a post-tag README polish commit (284baef304ce0ed785182635a4701895144a2673) for a cleaner project landing page. The archival release itself remains anchored to the final handoff tag.
QST pre-R2 history archive
Historical construction, acceptance, and cleanline records externalized before Stage R2 public product tree reset.