llama-sidecar v0.1.0: daemon + benchmarks + eval suite

Go daemon (cmd/llama-sidecar): per-agent llama-server process pool with
LRU eviction, OpenAI-compatible proxy, flag validation (Unsloth port),
deterministic hash-keyed sidecar reuse. Windows service support via
schtasks/NSSM with DETACHED_PROCESS, stdout pipe drain, and request-ctx
decoupled child lifetime.

Bug fixes (3b.1–3b5): -c flag drop from StripShadowingFlags, UTF-8 BOM
in JSON config, -fa → --flash-attn on default, child process exit after
one request (stdin devnull, stdout pipe, CREATE_NO_WINDOW → DETACHED,
context.Background for child lifetime, background reaper goroutine).

bench/: MTP on/off throughput sweep across 8 GGUFs via SSH+schtasks
automation to sam-desktop. Per-GGUF production flags from llama-swap
config with --ctx-size 32768 override.

eval/: accuracy benchmarks (MMLU 100q, GSM8K 50q, HumanEval 164) +
A/B model comparison (14 agent-typed prompts × 8 models). All scripts
resumable at individual question level.

94 Go tests, race detector clean.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

internal/server/admin.go

@@ -0,0 +1,42 @@
+package server
+
+import (
+	"net/http"
+	"time"
+
+	"github.com/indifferentketchup/llama-sidecar/internal/config"
+	"github.com/indifferentketchup/llama-sidecar/internal/pool"
+)
+
+func healthHandler(p *pool.Pool, cfg *config.Config, startedAt time.Time) http.HandlerFunc {
+	return func(w http.ResponseWriter, r *http.Request) {
+		sidecars := p.List()
+		writeJSON(w, http.StatusOK, map[string]any{
+			"status":         "ok",
+			"sidecars":       len(sidecars),
+			"max":            cfg.MaxSidecars,
+			"uptime_seconds": int(time.Since(startedAt).Seconds()),
+		})
+	}
+}
+
+func listSidecarsHandler(p *pool.Pool) http.HandlerFunc {
+	return func(w http.ResponseWriter, r *http.Request) {
+		writeJSON(w, http.StatusOK, p.List())
+	}
+}
+
+func deleteSidecarHandler(p *pool.Pool) http.HandlerFunc {
+	return func(w http.ResponseWriter, r *http.Request) {
+		hash := r.PathValue("hash")
+		if hash == "" {
+			writeJSON(w, http.StatusBadRequest, map[string]string{"error": "hash required"})
+			return
+		}
+		if err := p.Remove(hash); err != nil {
+			writeJSON(w, http.StatusNotFound, map[string]string{"error": err.Error()})
+			return
+		}
+		writeJSON(w, http.StatusOK, map[string]string{"status": "removed"})
+	}
+}