Dom 5c5ce747b0 feat(grounding): câblage Qwen3-VL-4B/vLLM (RPA_GROUNDING_ENGINE, défaut off) ...

Active via RPA_GROUNDING_ENGINE=qwen3vl_vllm (défaut OFF = legacy Qwen2.5-VL
inchangé, byte-identique). Mode qwen3vl : port 8001/Qwen3-VL-4B, prompt point
0-1, think=false, parse /1000 (dissout DETTE-006), method "grounding" gardée
(seuil 0.60), pas de fallback Ollama (abstention si vLLM down). Grounder validé
au bench Easily réel (0.933, ~1s/cas). TDD : 4 tests (normalisation 0-1000,
think=false, prompt fractions 0-1, gating score bas).

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

2026-06-13 08:39:29 +02:00

..

core

feat(dialog): R2 MVP P0 — DialogResolver + catalogue 10 entrées (flag OFF default)

2026-05-24 17:52:38 +02:00

__init__.py

chore: ajouter agent_v0/ au tracking git (était un repo embarqué)

2026-03-18 11:12:23 +01:00

agent_registry.py

feat(wp-c): méthode verify_token côté registre (patch 3, inerte)

2026-06-10 14:21:04 +02:00

api_stream.py

fix(security): bind 127.0.0.1 par défaut via RPA_BIND_HOST (plus de host=0.0.0.0 en dur)

2026-06-08 17:49:58 +02:00

audit_trail.py

feat: pipeline complet MACRO/MÉSO/MICRO — Critic, Observer, Policy, Recovery, Learning, Audit Trail, TaskPlanner

2026-04-09 21:03:25 +02:00

chat_interface.py

feat: Léa chat + IRBuilder enrichi (stratégies V4 complètes)

2026-04-10 09:01:13 +02:00

domain_context.py

feat(p1x): de-hardcode VLM models/endpoints to vlm_config (DGX-ready)

2026-06-03 14:06:03 +02:00

execution_plan_runner.py

fix: contrôle strict des étapes + routage par machine_id

2026-04-10 14:05:23 +02:00

live_session_manager.py

snapshot: WIP 5j replay reliability (B1 watchdog + dialog handlers + grounding drift)

2026-05-24 16:48:37 +02:00

loop_detector.py

feat(qw2): LoopDetector composite (screen_static + action_repeat + retry)

2026-05-05 23:09:43 +02:00

monitor_router.py

chore(qw): cleanup post-review (préfixes BUS, événements monitor, import io)

2026-05-06 00:08:22 +02:00

replay_engine.py

feat(p1): persist workflows and semantic learning artifacts

2026-06-02 16:20:38 +02:00

replay_failure_logger.py

chore: ajouter replay_failure_logger.py au tracking git

2026-04-12 10:35:51 +02:00

replay_learner.py

fix(agent): P0.6 guard human corrections

2026-05-24 21:07:12 +02:00

replay_memory.py

feat(p1): persist workflows and semantic learning artifacts

2026-06-02 16:20:38 +02:00

replay_verifier.py

feat(p1x): de-hardcode VLM models/endpoints to vlm_config (DGX-ready)

2026-06-03 14:06:03 +02:00

replay_watchdog.py

snapshot: WIP 5j replay reliability (B1 watchdog + dialog handlers + grounding drift)

2026-05-24 16:48:37 +02:00

resolve_engine.py

feat(grounding): câblage Qwen3-VL-4B/vLLM (RPA_GROUNDING_ENGINE, défaut off)

2026-06-13 08:39:29 +02:00

run_worker.py

fix(p0): secure agent revocation and R6 worker queue

2026-06-02 15:52:35 +02:00

safety_checks_provider.py

feat(p1x): de-hardcode VLM models/endpoints to vlm_config (DGX-ready)

2026-06-03 14:06:03 +02:00

session_worker.py

chore: ajouter agent_v0/ au tracking git (était un repo embarqué)

2026-03-18 11:12:23 +01:00

stream_processor.py

chore(dgx): snapshot consolidation WIP pour transfert poc DGX

2026-06-08 16:33:58 +02:00

task_planner.py

feat(p1x): de-hardcode VLM models/endpoints to vlm_config (DGX-ready)

2026-06-03 14:06:03 +02:00

visual_wait.py

chore: ajouter agent_v0/ au tracking git (était un repo embarqué)

2026-03-18 11:12:23 +01:00

vm_controller.py

chore: ajouter agent_v0/ au tracking git (était un repo embarqué)

2026-03-18 11:12:23 +01:00

worker_stream.py

fix(p0): secure agent revocation and R6 worker queue

2026-06-02 15:52:35 +02:00

workflow_replay.py

feat(p1): persist workflows and semantic learning artifacts

2026-06-02 16:20:38 +02:00