v0.4.1: native embedding CFFI — full pipeline working, daemon-wired, HITL bug fixed

- Native backend returns 768-dim vectors via llama.cpp / C wrapper (/usr/local/lib/libllama_wrap.so)
- Wired :native into embed-object dispatch and exported from passepartout package
- Model preloads at daemon startup with EMBEDDING_PROVIDER=native (~30s)
- Lazy loading via *embedding-backend* :native also works (first call ~45s)
- C wrapper bridges CFFI pointer params to llama.cpp struct-by-value API
- Correct struct layouts: llama_model_params(72B), llama_context_params(136B), llama_batch(56B)
- BERT pooling: llama_get_embeddings_seq, llama_tokenize takes vocab* not model*
- FiveAM tests pass: dimensions, self-similarity, semantic ranking
- Fixed pre-existing HITL crash: boundp guard for *hitl-pending* in core-loop-act
- Lazy load guard prevents double-load of native file in embedding-native-ensure-loaded
- ROADMAP: v0.4.0 items marked DONE, v0.4.1 native embedding updated with actual implementation

docs/CONTRIBUTING.org

@@ -36,7 +36,7 @@ Skills are the building blocks of Passepartout. They reside in the `skills/` dir
 A skill must define:
 1. *Trigger*: A lambda determining if the skill should activate based on the context.
 2. *Probabilistic Gate*: Optional. Generates a prompt for the LLM.
-3. *Deterministic Gate*: A hardcoded Lisp function that guarantees safety or executes side-effects (the "Bouncer" pattern).
+3. *Deterministic Gate*: A hardcoded Lisp function that guarantees safety or executes side-effects (the Dispatcher pattern).
 
 Example Registration:
 #+begin_src lisp