feat: run docs-mcp as a long-lived shared HTTP service

context-docs was previously spawned per call as a fresh stdio container,
which meant every MCP request paid full cold-start cost (embedding model
load + Chroma open) and concurrent clients raced for the same Chroma
writer. The 50+ orphan container build-up I saw during the publish audit
was the visible symptom.

This refactor runs docs-mcp as one long-lived service:

- compose: docs-mcp leaves the 'mcp' profile, gets container_name,
  restart: unless-stopped, healthcheck, and a host port (127.0.0.1:8776
  by default). Runs as the host UID/GID so bind mounts don't end up
  root-owned.
- docker image: adds mcp-proxy (0.12.0) and an entrypoint that fronts
  llms-txt-mcp's stdio as Streamable HTTP. Reads sources from a flat
  file mounted at /etc/context-kit/docs-sources.txt. Disables eager
  preindex by default; callers refresh on demand via the docs_refresh
  tool. Set CONTEXT_KIT_DOCS_PREINDEX=1 to restore eager behavior.
- bin/context-kit: 'start' brings up the docs service alongside SearXNG,
  generates the sources file from CONTEXT_KIT_DOCS_SOURCES, and waits
  for the HTTP endpoint to become ready (up to 180s for first-run model
  download). 'docs' still works for stdio-only clients but is now a
  thin mcp-proxy bridge onto the shared HTTP service. 'doctor' and
  'status' both surface the new endpoint.
- install snippets: context-docs is now 'type: remote'/'type: http'
  pointing at ${CONTEXT_KIT_DOCS_HTTP_URL}. HTTP-capable MCP clients
  bypass the bridge entirely. snippets/*.json and the install command
  output stay byte-identical.
- docs and .env.example updated for new vars (CONTEXT_KIT_DOCS_PORT,
  CONTEXT_KIT_DOCS_HTTP_URL, CONTEXT_KIT_DOCS_PREINDEX) and the new
  24h TTL default (down from 7d; the long-lived service makes shorter
  defaults cheap).

Verified end-to-end:
- compose config -q, bash -n, sh -n all clean
- HTTP /status returns 200
- stdio bridge returns initialize + tools/list with the same 3 tools
  (docs_sources, docs_refresh, docs_query)
- doctor passes all 10 checks including the new HTTP probe
- web-search and repomix MCP handshakes still work
- redaction-check clean
- install JSON valid for both targets + --absolute

docker/docs/entrypoint.sh

@@ -0,0 +1,53 @@
+#!/bin/sh
+# context-kit docs-mcp entrypoint.
+#
+# Bridges llms-txt-mcp (stdio-only) to Streamable HTTP via mcp-proxy so that
+# multiple clients share a single long-lived indexer instead of each spawning
+# their own container (and racing on the same Chroma store).
+#
+# Sources are read from $DOCS_MCP_SOURCES_FILE (one URL per line; `#` comments
+# and blank lines are allowed). Everything else is configured via env vars
+# with sensible defaults so this image works standalone too.
+
+set -eu
+
+sources_file="${DOCS_MCP_SOURCES_FILE:-/etc/context-kit/docs-sources.txt}"
+
+if [ ! -r "$sources_file" ]; then
+  echo "docs-mcp: sources file not readable: $sources_file" >&2
+  echo "docs-mcp: set DOCS_MCP_SOURCES_FILE or mount one at that path." >&2
+  exit 64
+fi
+
+# Strip comments and blank lines, then collapse whitespace into a flat list.
+sources=$(grep -vE '^[[:space:]]*(#|$)' "$sources_file" | tr -s '[:space:]' '\n' | grep -v '^$' || true)
+
+if [ -z "$sources" ]; then
+  echo "docs-mcp: no sources found in $sources_file after stripping comments/blanks" >&2
+  exit 64
+fi
+
+# By default llms-txt-mcp 0.2.0 re-embeds every source on launch (the actual
+# default is a background preindex, --no-preindex only disables the foreground
+# variant). On a long-lived container that just wastes ~5 min of CPU per
+# restart, so we disable BOTH and let the caller use `docs_refresh` on demand.
+# Set DOCS_MCP_PREINDEX=1 to restore the eager behavior.
+preindex_flag="--no-preindex --no-background-preindex"
+if [ "${DOCS_MCP_PREINDEX:-0}" = "1" ]; then
+  preindex_flag=""
+fi
+
+# shellcheck disable=SC2086  # intentional word-splitting on $sources / $preindex_flag
+exec mcp-proxy \
+  --host "${DOCS_MCP_HTTP_HOST:-0.0.0.0}" \
+  --port "${DOCS_MCP_HTTP_PORT:-8000}" \
+  --pass-environment \
+  --allow-origin "${DOCS_MCP_ALLOW_ORIGIN:-*}" \
+  -- \
+  llms-txt-mcp \
+    --store-path /data \
+    --ttl "${DOCS_MCP_TTL:-24h}" \
+    --max-get-bytes "${DOCS_MCP_MAX_GET_BYTES:-75000}" \
+    --embed-model "${DOCS_MCP_EMBED_MODEL:-BAAI/bge-small-en-v1.5}" \
+    $preindex_flag \
+    $sources