7 Commits

Author	SHA1	Message	Date
lama	3f4ff415bc	Harden first-use diagnostics and publish functional skill acceptance	2026-09-13 20:42:16 +08:00
lama	c704d115ec	Enhance agent runner and run agent scripts for model identity validation ... - Updated agent-runner documentation to clarify the requirement for exact model ID matching. - Implemented model identity checks in run_agent.py to ensure the response model matches the configured model. - Introduced a new script, run_paired_trial.py, to facilitate bounded A/B testing with a constrained skill-guidance protocol. - Added comprehensive tests for the paired trial CLI in test_run_paired_trial.py, covering various scenarios including credential handling, preflight checks, and model identity validation.	2026-09-07 12:28:16 +08:00
lama	20b8c3ef26	Add neutral trial control, durable budget gates and evidence failure checks	2026-09-06 19:12:56 +08:00
lama	8321c6831d	Add frozen paired-evaluation kit and honest calibration evidence	2026-09-06 18:25:07 +08:00
lama	bd3ba48821	fix: ground first-use reports in verified execution evidence	2026-09-06 17:05:51 +08:00
lama	0b3f4dd431	fix: harden installed harness and add evidence-driven learning path	2026-09-06 13:42:48 +08:00
lama	0edf6cb1c7	feat: build auditable research reproduction harness	2026-09-05 02:22:25 +08:00