Remove python implementation

.github/workflows/live.yml

@@ -3,7 +3,7 @@ name: Test - Live
 # Full agent×model end-to-end: spin a real llama.cpp server, build the per-agent
 # sandbox images on demand, and drive `agentcap run` against the server for each
 # agent (pi/hermes/goose). Heavy (GGUF download + image builds + CPU inference) —
-# this is agentcap's "live" tier, the Rust port of `linux-live-tests.yml`.
+# this is agentcap's "live" tier (real agent × model, end-to-end).
 
 on:
   push:
@@ -29,7 +29,7 @@ permissions:
 env:
   CARGO_TERM_COLOR: always
   RUST_BACKTRACE: "1"
-  # Pinned to match tests/conftest.py (the proven Python live setup).
+  # Pinned to the proven live setup (Qwen3-1.7B-Q8 on a CPU llama.cpp server).
   GGUF_REPO: Qwen/Qwen3-1.7B-GGUF
   GGUF_FILE: Qwen3-1.7B-Q8_0.gguf
   LLAMA_IMAGE: ghcr.io/ggml-org/llama.cpp:server-b9487

.github/workflows/test.yml

@@ -2,7 +2,7 @@ name: Test - Unit & Integration
 
 # Hermetic tests only: unit tests + the loopback proxy integration test. No
 # network, podman, or model server. The full agent×model end-to-end ("live")
-# tests are a separate, resource-heavy category — see `linux-live-tests.yml`.
+# tests are a separate, resource-heavy category — see `live.yml`.
 
 on:
   pull_request:

.gitignore

@@ -1,19 +1,10 @@
-__pycache__/
-*.pyc
-.venv/
-.venv*/
-build/
-dist/
-*.egg-info/
 runs/
 .agentcap/
 examples/*/sandbox/
-.pytest_cache/
-.ruff_cache/
 .vscode/
 .idea/
 .DS_Store
 
 # Rust
 /target/
-
+dist/