diff --git a/.claude-plugin/marketplace.json b/.claude-plugin/marketplace.json
new file mode 100644
index 00000000..24d16eee
--- /dev/null
+++ b/.claude-plugin/marketplace.json
@@ -0,0 +1,14 @@
+{
+  "name": "haiku-rag",
+  "description": "The haiku.rag knowledge base as Claude Code tools and a skill.",
+  "owner": {
+    "name": "Yiorgis Gozadinos"
+  },
+  "plugins": [
+    {
+      "name": "haiku-rag",
+      "source": "./claude-plugin",
+      "description": "Search, read and question your haiku.rag knowledge base from Claude Code."
+    }
+  ]
+}
diff --git a/CHANGELOG.md b/CHANGELOG.md
index 7a22246e..41411c13 100644
--- a/CHANGELOG.md
+++ b/CHANGELOG.md
@@ -4,6 +4,9 @@
 
 ### Added
 
+- Claude Code plugin under `claude-plugin/`: the server configuration and the
+  `haiku-rag` skill. `claude plugin marketplace add ggozad/haiku.rag`, then
+  `claude plugin install haiku-rag`.
 - `haiku-rag mcp --no-agents` leaves `ask_question` and `analyze`
   unregistered. `create_mcp_server(agents=)`, `HaikuRAGApp.run_mcp(agents=)`.
 - MCP tools `get_document_outline` (heading tree with page numbers) and
diff --git a/README.md b/README.md
index 9016e322..fb3c8d5b 100644
--- a/README.md
+++ b/README.md
@@ -110,12 +110,19 @@ For direct agent composition, see the [capabilities documentation](https://ggoza
 
 ## MCP Server
 
-Use with AI assistants like Claude Desktop:
+Use with AI assistants like Claude Code and Claude Desktop:
 
 ```bash
 haiku-rag mcp --stdio
 ```
 
+In Claude Code, install the plugin, which registers the server and a skill:
+
+```bash
+claude plugin marketplace add ggozad/haiku.rag
+claude plugin install haiku-rag
+```
+
 Add to your Claude Desktop configuration:
 
 ```json
diff --git a/claude-plugin/.claude-plugin/plugin.json b/claude-plugin/.claude-plugin/plugin.json
new file mode 100644
index 00000000..93282056
--- /dev/null
+++ b/claude-plugin/.claude-plugin/plugin.json
@@ -0,0 +1,12 @@
+{
+  "name": "haiku-rag",
+  "version": "0.82.1",
+  "description": "Search, read and question your haiku.rag knowledge base from Claude Code.",
+  "author": {
+    "name": "Yiorgis Gozadinos",
+    "email": "ggozadinos@gmail.com"
+  },
+  "homepage": "https://ggozad.github.io/haiku.rag/mcp/",
+  "repository": "https://github.com/ggozad/haiku.rag",
+  "license": "MIT"
+}
diff --git a/claude-plugin/.mcp.json b/claude-plugin/.mcp.json
new file mode 100644
index 00000000..f6fb25f9
--- /dev/null
+++ b/claude-plugin/.mcp.json
@@ -0,0 +1,8 @@
+{
+  "mcpServers": {
+    "haiku-rag": {
+      "command": "haiku-rag",
+      "args": ["mcp", "--stdio"]
+    }
+  }
+}
diff --git a/claude-plugin/skills/haiku-rag/SKILL.md b/claude-plugin/skills/haiku-rag/SKILL.md
new file mode 100644
index 00000000..f93631a3
--- /dev/null
+++ b/claude-plugin/skills/haiku-rag/SKILL.md
@@ -0,0 +1,65 @@
+---
+name: haiku-rag
+description: Search, read and question the user's haiku.rag knowledge base
+  through the haiku-rag MCP tools. Use whenever a request could be answered
+  from the user's ingested documents, when asked to find, look up, check or
+  cite something in their documents or knowledge base, or when the question is
+  about the user's own material rather than general knowledge.
+allowed-tools:
+  - mcp__plugin_haiku-rag_haiku-rag__search_documents
+  - mcp__plugin_haiku-rag_haiku-rag__search_documents_by_image
+  - mcp__plugin_haiku-rag_haiku-rag__get_document
+  - mcp__plugin_haiku-rag_haiku-rag__get_document_outline
+  - mcp__plugin_haiku-rag_haiku-rag__get_document_section
+  - mcp__plugin_haiku-rag_haiku-rag__list_documents
+  - mcp__plugin_haiku-rag_haiku-rag__ask_question
+  - mcp__plugin_haiku-rag_haiku-rag__analyze
+---
+
+# Working with the knowledge base
+
+Check the knowledge base before answering from memory whenever the question
+could be about the user's documents. Say so when it has nothing relevant.
+
+## Find
+
+`search_documents` is the first call. Results come best first with the document
+title, section headings and the matching passage. `filter` restricts which
+documents are searched, `limit` how many results come back. If it misses,
+rephrase once or narrow with a filter before concluding the material is not
+there.
+
+## Read
+
+Every search result shows its `Document ID` (and `Collection` when there are
+several); pass them to the read tools. `get_document` returns a document's
+whole text in reading order. For a long one, `get_document_outline` gives the
+heading tree with page numbers and `get_document_section` the text of one
+section, subsections included.
+
+## Answer or compute
+
+`ask_question` runs the RAG agent on the server and returns an answer with
+citations; use it when the user wants an answer rather than material.
+`analyze` runs code in a sandbox over the documents; use it for counting,
+aggregation, comparison across many documents or computation over tables. Both
+cost a model call and are slower than a search.
+
+## Explore
+
+`list_documents` shows what is stored: titles, URIs and metadata. It is how you
+learn what a filter can match.
+
+## Filters
+
+A SQL WHERE clause over the document columns `id`, `uri`, `title`,
+`created_at`, `updated_at`, `metadata`. `metadata` is a JSON string, so match
+it with LIKE: `metadata LIKE '%"author": "Smith"%'`. Also `uri LIKE '%.pdf'`,
+`title = 'Q3 report'`.
+
+## Results and citations
+
+Rank is the signal; scores are not comparable across queries and are never
+confidence. Cite the document title or URI, the section heading and page
+numbers when present. When results carry `source`, the server covers several
+collections: name it, and pass `sources` to search a subset.
diff --git a/docs/mcp.md b/docs/mcp.md
index 59093b3b..cbf08999 100644
--- a/docs/mcp.md
+++ b/docs/mcp.md
@@ -40,6 +40,25 @@ document. A name the server does not cover is an error.
 `haiku-rag --db-name NAME mcp` serves one. See
 [Multiple Databases](configuration/storage.md#multiple-databases).
 
+## Claude Code
+
+The repository ships a plugin that registers the server and a skill telling
+Claude when and how to use it:
+
+```bash
+claude plugin marketplace add ggozad/haiku.rag
+claude plugin install haiku-rag
+```
+
+The plugin runs `haiku-rag mcp --stdio`, so `haiku-rag` must be on the PATH
+and the configuration decides the database. The skill pre-approves every tool
+and is also invocable as `/haiku-rag`. To register the server without the
+plugin:
+
+```bash
+claude mcp add haiku-rag -- haiku-rag mcp --stdio
+```
+
 ## Claude Desktop Integration
 
 Add to your Claude Desktop configuration (`claude_desktop_config.json`):
diff --git a/scripts/bump_version.py b/scripts/bump_version.py
index 5c5a1e2e..9593cd89 100755
--- a/scripts/bump_version.py
+++ b/scripts/bump_version.py
@@ -2,7 +2,8 @@
 """
 Version bumping script for haiku.rag workspace.
 
-Updates version in all pyproject.toml files and CHANGELOG.md.
+Updates version in all pyproject.toml files, the Claude Code plugin manifest
+and CHANGELOG.md.
 """
 
 import re
@@ -54,6 +55,19 @@ def update_example_dependencies(file_path: Path, new_version: str) -> None:
     print(f"✓ Updated example dependencies in {file_path.relative_to(Path.cwd())}")
 
 
+def update_plugin_version(file_path: Path, new_version: str) -> None:
+    """Update the version in the Claude Code plugin manifest."""
+    content = file_path.read_text()
+    updated = re.sub(
+        r'^(\s*"version": )"[^"]+"',
+        rf'\1"{new_version}"',
+        content,
+        flags=re.MULTILINE,
+    )
+    file_path.write_text(updated)
+    print(f"✓ Updated {file_path.relative_to(Path.cwd())}")
+
+
 def update_changelog(changelog_path: Path, new_version: str) -> None:
     """Update CHANGELOG.md with new version."""
     content = changelog_path.read_text()
@@ -122,10 +136,13 @@ def main():
         root / "app" / "backend" / "pyproject.toml",
     ]
 
+    plugin_file = root / "claude-plugin" / ".claude-plugin" / "plugin.json"
     changelog_file = root / "CHANGELOG.md"
 
     # Check all files exist
-    for file in pyproject_files + example_pyproject_files + [changelog_file]:
+    for file in (
+        pyproject_files + example_pyproject_files + [plugin_file, changelog_file]
+    ):
         if not file.exists():
             print(f"Error: {file} not found")
             sys.exit(1)
@@ -155,6 +172,8 @@ def main():
     for file in example_pyproject_files:
         update_example_dependencies(file, new_version)
 
+    update_plugin_version(plugin_file, new_version)
+
     # Update CHANGELOG.md
     update_changelog(changelog_file, new_version)
 
diff --git a/tests/test_bump_version.py b/tests/test_bump_version.py
new file mode 100644
index 00000000..4f8fcf4f
--- /dev/null
+++ b/tests/test_bump_version.py
@@ -0,0 +1,38 @@
+import importlib.util
+import json
+from pathlib import Path
+
+_spec = importlib.util.spec_from_file_location(
+    "bump_version", Path(__file__).resolve().parents[1] / "scripts" / "bump_version.py"
+)
+assert _spec is not None and _spec.loader is not None
+bump_version = importlib.util.module_from_spec(_spec)
+_spec.loader.exec_module(bump_version)
+
+
+def test_update_plugin_version_rewrites_only_the_version_field(tmp_path, monkeypatch):
+    monkeypatch.chdir(tmp_path)
+    manifest = tmp_path / "plugin.json"
+    manifest.write_text(
+        '{\n  "name": "haiku-rag",\n  "version": "0.1.0",\n  "license": "MIT"\n}\n'
+    )
+
+    bump_version.update_plugin_version(manifest, "0.2.0")
+
+    assert json.loads(manifest.read_text()) == {
+        "name": "haiku-rag",
+        "version": "0.2.0",
+        "license": "MIT",
+    }
+    assert manifest.read_text().endswith("}\n")
+
+
+def test_the_shipped_plugin_manifest_carries_the_package_version():
+    root = Path(__file__).resolve().parents[1]
+    plugin = json.loads(
+        (root / "claude-plugin" / ".claude-plugin" / "plugin.json").read_text()
+    )
+
+    assert plugin["version"] == bump_version.get_current_version(
+        root / "haiku_rag_slim" / "pyproject.toml"
+    )
diff --git a/tests/test_mcp.py b/tests/test_mcp.py
index 7bf3183d..4f1160b8 100644
--- a/tests/test_mcp.py
+++ b/tests/test_mcp.py
@@ -1,4 +1,5 @@
 import logging
+from pathlib import Path
 from types import SimpleNamespace
 
 import pytest
@@ -1122,6 +1123,48 @@ class TestMCPErrorContract:
         )
 
 
+class TestClaudeCodePlugin:
+    """The plugin under claude-plugin/ points at the server this module builds."""
+
+    root = Path(__file__).resolve().parents[1]
+
+    def test_the_manifests_name_the_plugin_and_its_server(self):
+        import json
+
+        plugin = json.loads(
+            (self.root / "claude-plugin/.claude-plugin/plugin.json").read_text()
+        )
+        marketplace = json.loads(
+            (self.root / ".claude-plugin/marketplace.json").read_text()
+        )
+        servers = json.loads((self.root / "claude-plugin/.mcp.json").read_text())
+
+        assert plugin["name"] == "haiku-rag"
+        assert plugin["description"]
+        [entry] = marketplace["plugins"]
+        assert entry["name"] == plugin["name"]
+        assert entry["source"] == "./claude-plugin"
+        assert servers["mcpServers"]["haiku-rag"]["args"] == ["mcp", "--stdio"]
+
+    @pytest.mark.asyncio
+    async def test_the_skill_pre_approves_every_tool_the_server_registers(
+        self, mcp_db, multimodal_embedder
+    ):
+        import yaml
+
+        text = (self.root / "claude-plugin/skills/haiku-rag/SKILL.md").read_text()
+        _, frontmatter, _ = text.split("---", 2)
+        skill = yaml.safe_load(frontmatter)
+        prefix = "mcp__plugin_haiku-rag_haiku-rag__"
+
+        assert skill["name"] == "haiku-rag"
+        assert skill["description"]
+        assert all(tool.startswith(prefix) for tool in skill["allowed-tools"])
+        approved = {tool.removeprefix(prefix) for tool in skill["allowed-tools"]}
+        registered = {t.name for t in await create_mcp_server(mcp_db).list_tools()}
+        assert approved == registered
+
+
 class TestMCPClientLifetime:
     @pytest.mark.asyncio
     async def test_tool_calls_share_one_database_open(self, mcp_db, monkeypatch):