Marius CiepluchandClaude Fable 5.1 bacd7a1e3f feat(hunt): proposals and the acceptance rule (5.1, 5.2) ...

propose.py: one change per round (instruction, demonstrations or one mask), examples built from
compile folds only with a provenance record of the change ids used, a registry for GEPA, MIPROv2
and BootstrapFewShot (MIPROv2 without the data-aware proposer so labels are never serialised),
compile-fold scoring with Wilson and the repository-paired bootstrap, and a compile-only history
the next round may read. accept.py: the rule as independent clauses (both folds non-negative with
one strictly positive, per-family regression budget on held-out recall, tie tolerance, no
acceptance without an interval, instruction cap, one change only), the comparison recorded in the
candidate artifact's configurations, rejections logged without held-out numbers. 48 fixture tests.
Known gap for live rounds: the optimisers and propose._score call the program directly, so a
tool-using program must be scored through the runner with its tools bound per checkout (6.1).
Implemented by Codex (gpt-6-astra), gated here.

Co-Authored-By: Claude Fable 5.1 <[email protected]>

2026-10-13 20:22:22 +02:00

..

design.md

spec(hunter-learning-loop): a pattern engine's miss is never evidence

2026-10-13 13:01:32 +02:00

requirements.md

spec(hunter-learning-loop): requirements

2026-10-13 08:47:51 +02:00

research.md

research: Semgrep CE and Opengrep as the quick tier, measured

2026-10-13 12:21:02 +02:00

spec.json

spec(hunter-learning-loop): implementation plan

2026-10-13 12:15:45 +02:00

tasks.md

feat(hunt): proposals and the acceptance rule (5.1, 5.2)

2026-10-13 20:22:22 +02:00