karst 0.1.0__tar.gz

karst-0.1.0/LICENSE

@@ -0,0 +1,201 @@
+                                 Apache License
+                           Version 2.0, January 2004
+                        http://www.apache.org/licenses/
+
+   TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION
+
+   1. Definitions.
+
+      "License" shall mean the terms and conditions for use, reproduction,
+      and distribution as defined by Sections 1 through 9 of this document.
+
+      "Licensor" shall mean the copyright owner or entity authorized by
+      the copyright owner that is granting the License.
+
+      "Legal Entity" shall mean the union of the acting entity and all
+      other entities that control, are controlled by, or are under common
+      control with that entity. For the purposes of this definition,
+      "control" means (i) the power, direct or indirect, to cause the
+      direction or management of such entity, whether by contract or
+      otherwise, or (ii) ownership of fifty percent (50%) or more of the
+      outstanding shares, or (iii) beneficial ownership of such entity.
+
+      "You" (or "Your") shall mean an individual or Legal Entity
+      exercising permissions granted by this License.
+
+      "Source" form shall mean the preferred form for making modifications,
+      including but not limited to software source code, documentation
+      source, and configuration files.
+
+      "Object" form shall mean any form resulting from mechanical
+      transformation or translation of a Source form, including but
+      not limited to compiled object code, generated documentation,
+      and conversions to other media types.
+
+      "Work" shall mean the work of authorship, whether in Source or
+      Object form, made available under the License, as indicated by a
+      copyright notice that is included in or attached to the work
+      (an example is provided in the Appendix below).
+
+      "Derivative Works" shall mean any work, whether in Source or Object
+      form, that is based on (or derived from) the Work and for which the
+      editorial revisions, annotations, elaborations, or other modifications
+      represent, as a whole, an original work of authorship. For the purposes
+      of this License, Derivative Works shall not include works that remain
+      separable from, or merely link (or bind by name) to the interfaces of,
+      the Work and Derivative Works thereof.
+
+      "Contribution" shall mean any work of authorship, including
+      the original version of the Work and any modifications or additions
+      to that Work or Derivative Works thereof, that is intentionally
+      submitted to Licensor for inclusion in the Work by the copyright owner
+      or by an individual or Legal Entity authorized to submit on behalf of
+      the copyright owner. For the purposes of this definition, "submitted"
+      means any form of electronic, verbal, or written communication sent
+      to the Licensor or its representatives, including but not limited to
+      communication on electronic mailing lists, source code control systems,
+      and issue tracking systems that are managed by, or on behalf of, the
+      Licensor for the purpose of discussing and improving the Work, but
+      excluding communication that is conspicuously marked or otherwise
+      designated in writing by the copyright owner as "Not a Contribution."
+
+      "Contributor" shall mean Licensor and any individual or Legal Entity
+      on behalf of whom a Contribution has been received by Licensor and
+      subsequently incorporated within the Work.
+
+   2. Grant of Copyright License. Subject to the terms and conditions of
+      this License, each Contributor hereby grants to You a perpetual,
+      worldwide, non-exclusive, no-charge, royalty-free, irrevocable
+      copyright license to reproduce, prepare Derivative Works of,
+      publicly display, publicly perform, sublicense, and distribute the
+      Work and such Derivative Works in Source or Object form.
+
+   3. Grant of Patent License. Subject to the terms and conditions of
+      this License, each Contributor hereby grants to You a perpetual,
+      worldwide, non-exclusive, no-charge, royalty-free, irrevocable
+      (except as stated in this section) patent license to make, have made,
+      use, offer to sell, sell, import, and otherwise transfer the Work,
+      where such license applies only to those patent claims licensable
+      by such Contributor that are necessarily infringed by their
+      Contribution(s) alone or by combination of their Contribution(s)
+      with the Work to which such Contribution(s) was submitted. If You
+      institute patent litigation against any entity (including a
+      cross-claim or counterclaim in a lawsuit) alleging that the Work
+      or a Contribution incorporated within the Work constitutes direct
+      or contributory patent infringement, then any patent licenses
+      granted to You under this License for that Work shall terminate
+      as of the date such litigation is filed.
+
+   4. Redistribution. You may reproduce and distribute copies of the
+      Work or Derivative Works thereof in any medium, with or without
+      modifications, and in Source or Object form, provided that You
+      meet the following conditions:
+
+      (a) You must give any other recipients of the Work or Derivative
+          Works a copy of this License; and
+
+      (b) You must cause any modified files to carry prominent notices
+          stating that You changed the files; and
+
+      (c) You must retain, in the Source form of any Derivative Works
+          that You distribute, all copyright, patent, trademark, and
+          attribution notices from the Source form of the Work,
+          excluding those notices that do not pertain to any part of
+          the Derivative Works; and
+
+      (d) If the Work includes a "NOTICE" text file as part of its
+          distribution, then any Derivative Works that You distribute must
+          include a readable copy of the attribution notices contained
+          within such NOTICE file, excluding those notices that do not
+          pertain to any part of the Derivative Works, in at least one
+          of the following places: within a NOTICE text file distributed
+          as part of the Derivative Works; within the Source form or
+          documentation, if provided along with the Derivative Works; or,
+          within a display generated by the Derivative Works, if and
+          wherever such third-party notices normally appear. The contents
+          of the NOTICE file are for informational purposes only and do
+          not modify the License. You may add Your own attribution notices
+          within Derivative Works that You distribute, alongside or as an
+          addendum to the NOTICE text from the Work, provided that such
+          additional attribution notices cannot be construed as modifying
+          the License.
+
+      You may add Your own copyright statement to Your modifications and
+      may provide additional or different license terms and conditions
+      for use, reproduction, or distribution of Your modifications, or
+      for any such Derivative Works as a whole, provided Your use,
+      reproduction, and distribution of the Work otherwise complies with
+      the conditions stated in this License.
+
+   5. Submission of Contributions. Unless You explicitly state otherwise,
+      any Contribution intentionally submitted for inclusion in the Work
+      by You to the Licensor shall be under the terms and conditions of
+      this License, without any additional terms or conditions.
+      Notwithstanding the above, nothing herein shall supersede or modify
+      the terms of any separate license agreement you may have executed
+      with Licensor regarding such Contributions.
+
+   6. Trademarks. This License does not grant permission to use the trade
+      names, trademarks, service marks, or product names of the Licensor,
+      except as required for reasonable and customary use in describing the
+      origin of the Work and reproducing the content of the NOTICE file.
+
+   7. Disclaimer of Warranty. Unless required by applicable law or
+      agreed to in writing, Licensor provides the Work (and each
+      Contributor provides its Contributions) on an "AS IS" BASIS,
+      WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
+      implied, including, without limitation, any warranties or conditions
+      of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A
+      PARTICULAR PURPOSE. You are solely responsible for determining the
+      appropriateness of using or redistributing the Work and assume any
+      risks associated with Your exercise of permissions under this License.
+
+   8. Limitation of Liability. In no event and under no legal theory,
+      whether in tort (including negligence), contract, or otherwise,
+      unless required by applicable law (such as deliberate and grossly
+      negligent acts) or agreed to in writing, shall any Contributor be
+      liable to You for damages, including any direct, indirect, special,
+      incidental, or consequential damages of any character arising as a
+      result of this License or out of the use or inability to use the
+      Work (including but not limited to damages for loss of goodwill,
+      work stoppage, computer failure or malfunction, or any and all
+      other commercial damages or losses), even if such Contributor
+      has been advised of the possibility of such damages.
+
+   9. Accepting Warranty or Additional Liability. While redistributing
+      the Work or Derivative Works thereof, You may choose to offer,
+      and charge a fee for, acceptance of support, warranty, indemnity,
+      or other liability obligations and/or rights consistent with this
+      License. However, in accepting such obligations, You may act only
+      on Your own behalf and on Your sole responsibility, not on behalf
+      of any other Contributor, and only if You agree to indemnify,
+      defend, and hold each Contributor harmless for any liability
+      incurred by, or claims asserted against, such Contributor by reason
+      of your accepting any such warranty or additional liability.
+
+   END OF TERMS AND CONDITIONS
+
+   APPENDIX: How to apply the Apache License to your work.
+
+      To apply the Apache License to your work, attach the following
+      boilerplate notice, with the fields enclosed by brackets "[]"
+      replaced with your own identifying information. (Don't include
+      the brackets!)  The text should be enclosed in the appropriate
+      comment syntax for the file format. We also recommend that a
+      file or class name and description of purpose be included on the
+      same "printed page" as the copyright notice for easier
+      identification within third-party archives.
+
+   Copyright 2026 karst
+
+   Licensed under the Apache License, Version 2.0 (the "License");
+   you may not use this file except in compliance with the License.
+   You may obtain a copy of the License at
+
+       http://www.apache.org/licenses/LICENSE-2.0
+
+   Unless required by applicable law or agreed to in writing, software
+   distributed under the License is distributed on an "AS IS" BASIS,
+   WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+   See the License for the specific language governing permissions and
+   limitations under the License.

karst-0.1.0/PKG-INFO

@@ -0,0 +1,151 @@
+Metadata-Version: 2.4
+Name: karst
+Version: 0.1.0
+Summary: Code context for AI dev tools — graph-grounded, pack-scoped retrieval over MCP. 60% fewer tokens, audit-grade citations.
+Author: karst
+License: Apache-2.0
+Project-URL: Homepage, https://github.com/Moin105/upgraded-garbanzo
+Project-URL: Repository, https://github.com/Moin105/upgraded-garbanzo
+Project-URL: Issues, https://github.com/Moin105/upgraded-garbanzo/issues
+Keywords: mcp,model-context-protocol,code-search,rag,graphrag,tree-sitter,code-intelligence,ai,llm,cursor,claude,embeddings,code-review,impact-analysis
+Classifier: Development Status :: 4 - Beta
+Classifier: Intended Audience :: Developers
+Classifier: License :: OSI Approved :: Apache Software License
+Classifier: Operating System :: OS Independent
+Classifier: Programming Language :: Python :: 3
+Classifier: Programming Language :: Python :: 3.10
+Classifier: Programming Language :: Python :: 3.11
+Classifier: Programming Language :: Python :: 3.12
+Classifier: Programming Language :: Python :: 3.13
+Classifier: Topic :: Software Development :: Libraries
+Classifier: Topic :: Software Development :: Quality Assurance
+Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
+Requires-Python: >=3.10
+Description-Content-Type: text/markdown
+License-File: LICENSE
+Requires-Dist: tree-sitter>=0.23
+Requires-Dist: tree-sitter-language-pack>=0.7.3
+Requires-Dist: pathspec>=0.12
+Requires-Dist: fastembed>=0.4
+Requires-Dist: qdrant-client>=1.12
+Requires-Dist: unidiff>=0.7.5
+Requires-Dist: networkx>=3.2
+Requires-Dist: mcp>=1.2
+Provides-Extra: anthropic
+Requires-Dist: anthropic>=0.39; extra == "anthropic"
+Provides-Extra: openai
+Requires-Dist: openai>=1.50; extra == "openai"
+Provides-Extra: llm
+Requires-Dist: anthropic>=0.39; extra == "llm"
+Requires-Dist: openai>=1.50; extra == "llm"
+Provides-Extra: dev
+Requires-Dist: pytest>=8; extra == "dev"
+Dynamic: license-file
+
+# karst
+
+**Code context for AI dev tools.** karst sits between your repo and any AI
+tool — Cursor, Claude Desktop, a custom agent — and feeds it the *right* slice
+of the codebase: graph-grounded, pack-scoped, and cited to `file:line`. The
+result is ~60% fewer input tokens per question, answers you can verify, and a
+blast-radius check before you change anything.
+
+It runs **locally**, returns **context (not answers)** over **MCP**, and never
+calls an LLM itself — so you don't give karst an API key. Your IDE already has
+the model; karst just makes what it reads sharp and cheap.
+
+```
+pip install karst
+```
+
+## Why
+
+Most "chat with your codebase" tools dump tens of thousands of vaguely-related
+tokens into the model on every question. You can't see what was loaded, you
+can't scope it, and the bill arrives at the end of the month. karst inverts
+that:
+
+- **Scopes** — pack-filtered retrieval reads ~200 chunks, not 5,000.
+- **Cites** — every chunk carries an exact `file:line`. Verify, don't trust.
+- **Predicts** — a real call/import graph answers "what else breaks if I change
+  this?" — which embeddings alone can't.
+
+Measured on a real 246-file NestJS + Next.js repo: 906 chunks indexed, re-index
+**343s → 2.3s** incremental, **~$0.019** per question on Sonnet 4.6 (shown
+*before* the call), **60%** fewer tokens with packs attached.
+
+## Quickstart (CLI)
+
+```bash
+# 1. index a repo (incremental + cached after the first run)
+karst index ./my-repo
+
+# 2. build the call/import graph (enables impact analysis)
+karst graph-index ./my-repo
+
+# 3. auto-suggest context packs and tag the index
+karst packs --storage ~/.karst/indexes/my-repo \
+  suggest ./my-repo --apply --retag
+
+# 4. ask — retrieval is pack-scoped and the token cost is printed
+karst ask "How does checkout charge the user?" \
+  --storage ~/.karst/indexes/my-repo
+
+# what breaks if I change a function?
+karst impact --target checkout \
+  --graph-path ~/.karst/indexes/my-repo/graph.pkl
+
+# review a diff with severity-tagged, cited findings
+karst review --staged --storage ~/.karst/indexes/my-repo
+```
+
+`karst ask` needs an LLM key (`ANTHROPIC_API_KEY` or `OPENAI_API_KEY`), or pass
+`--no-llm` to get the raw cited chunks. The **MCP server below needs no key** —
+your IDE supplies the model.
+
+## Use it from your IDE (MCP)
+
+karst ships an MCP server (`karst-mcp`) exposing five tools — `search_code`,
+`find_impact`, `list_packs`, `index_status`, `index_repository` — over stdio.
+
+**Claude Desktop** (`claude_desktop_config.json`) or **Cursor**
+(`.cursor/mcp.json`):
+
+```json
+{
+  "mcpServers": {
+    "karst": { "command": "karst-mcp" }
+  }
+}
+```
+
+Restart the host, then ask normally — it calls karst's tools when useful and
+gets back scoped, cited context. Full setup, including the `python -m
+karst.mcp_server` fallback, is in [docs/MCP.md](docs/MCP.md).
+
+## How it works
+
+1. **Index** — tree-sitter splits every function, class and method into an
+   AST-aware chunk (Python, JS, TS, Go, Rust, Java); chunks are embedded into a
+   local Qdrant store. Incremental: a SHA manifest + embedding cache skip
+   unchanged files.
+2. **Graph** — a NetworkX knowledge graph of `CALLS` / `IMPORTS` / `CONTAINS`
+   edges powers impact analysis ("what depends on this?").
+3. **Pack** — related files become named, attachable context packs (`auth`,
+   `billing`). A query loads only its pack.
+4. **Serve** — the MCP server returns ranked, `file:line`-cited chunks; your
+   host's model reasons over them.
+
+Everything is local and offline-capable (FastEmbed/ONNX embeddings, Qdrant
+local mode, sqlite caches — no Docker, no daemon).
+
+## Status
+
+Live: AST chunking (6 languages), call/import graph + impact analysis,
+pack-scoped retrieval, token + cost meter, incremental indexing + embedding
+cache, diff code review, and the MCP server. Coming next: hosted indexing,
+team-shared pack libraries, a GitHub PR review bot.
+
+## License
+
+Apache-2.0. See [LICENSE](LICENSE).

karst-0.1.0/README.md

@@ -0,0 +1,107 @@
+# karst
+
+**Code context for AI dev tools.** karst sits between your repo and any AI
+tool — Cursor, Claude Desktop, a custom agent — and feeds it the *right* slice
+of the codebase: graph-grounded, pack-scoped, and cited to `file:line`. The
+result is ~60% fewer input tokens per question, answers you can verify, and a
+blast-radius check before you change anything.
+
+It runs **locally**, returns **context (not answers)** over **MCP**, and never
+calls an LLM itself — so you don't give karst an API key. Your IDE already has
+the model; karst just makes what it reads sharp and cheap.
+
+```
+pip install karst
+```
+
+## Why
+
+Most "chat with your codebase" tools dump tens of thousands of vaguely-related
+tokens into the model on every question. You can't see what was loaded, you
+can't scope it, and the bill arrives at the end of the month. karst inverts
+that:
+
+- **Scopes** — pack-filtered retrieval reads ~200 chunks, not 5,000.
+- **Cites** — every chunk carries an exact `file:line`. Verify, don't trust.
+- **Predicts** — a real call/import graph answers "what else breaks if I change
+  this?" — which embeddings alone can't.
+
+Measured on a real 246-file NestJS + Next.js repo: 906 chunks indexed, re-index
+**343s → 2.3s** incremental, **~$0.019** per question on Sonnet 4.6 (shown
+*before* the call), **60%** fewer tokens with packs attached.
+
+## Quickstart (CLI)
+
+```bash
+# 1. index a repo (incremental + cached after the first run)
+karst index ./my-repo
+
+# 2. build the call/import graph (enables impact analysis)
+karst graph-index ./my-repo
+
+# 3. auto-suggest context packs and tag the index
+karst packs --storage ~/.karst/indexes/my-repo \
+  suggest ./my-repo --apply --retag
+
+# 4. ask — retrieval is pack-scoped and the token cost is printed
+karst ask "How does checkout charge the user?" \
+  --storage ~/.karst/indexes/my-repo
+
+# what breaks if I change a function?
+karst impact --target checkout \
+  --graph-path ~/.karst/indexes/my-repo/graph.pkl
+
+# review a diff with severity-tagged, cited findings
+karst review --staged --storage ~/.karst/indexes/my-repo
+```
+
+`karst ask` needs an LLM key (`ANTHROPIC_API_KEY` or `OPENAI_API_KEY`), or pass
+`--no-llm` to get the raw cited chunks. The **MCP server below needs no key** —
+your IDE supplies the model.
+
+## Use it from your IDE (MCP)
+
+karst ships an MCP server (`karst-mcp`) exposing five tools — `search_code`,
+`find_impact`, `list_packs`, `index_status`, `index_repository` — over stdio.
+
+**Claude Desktop** (`claude_desktop_config.json`) or **Cursor**
+(`.cursor/mcp.json`):
+
+```json
+{
+  "mcpServers": {
+    "karst": { "command": "karst-mcp" }
+  }
+}
+```
+
+Restart the host, then ask normally — it calls karst's tools when useful and
+gets back scoped, cited context. Full setup, including the `python -m
+karst.mcp_server` fallback, is in [docs/MCP.md](docs/MCP.md).
+
+## How it works
+
+1. **Index** — tree-sitter splits every function, class and method into an
+   AST-aware chunk (Python, JS, TS, Go, Rust, Java); chunks are embedded into a
+   local Qdrant store. Incremental: a SHA manifest + embedding cache skip
+   unchanged files.
+2. **Graph** — a NetworkX knowledge graph of `CALLS` / `IMPORTS` / `CONTAINS`
+   edges powers impact analysis ("what depends on this?").
+3. **Pack** — related files become named, attachable context packs (`auth`,
+   `billing`). A query loads only its pack.
+4. **Serve** — the MCP server returns ranked, `file:line`-cited chunks; your
+   host's model reasons over them.
+
+Everything is local and offline-capable (FastEmbed/ONNX embeddings, Qdrant
+local mode, sqlite caches — no Docker, no daemon).
+
+## Status
+
+Live: AST chunking (6 languages), call/import graph + impact analysis,
+pack-scoped retrieval, token + cost meter, incremental indexing + embedding
+cache, diff code review, and the MCP server. Coming next: hosted indexing,
+team-shared pack libraries, a GitHub PR review bot.
+
+## License
+
+Apache-2.0. See [LICENSE](LICENSE).

karst-0.1.0/karst/__init__.py

@@ -0,0 +1 @@
+__version__ = "0.1.0"

karst-0.1.0/karst/analyze.py

@@ -0,0 +1,39 @@
+"""End-to-end analyze pipeline: walk → parse → chunk.
+
+Holds the public surface that the CLI (and later, agents) will call.
+"""
+
+from __future__ import annotations
+
+from collections.abc import Iterator
+from dataclasses import dataclass
+from pathlib import Path
+
+from .chunker import chunk_file
+from .models import Chunk
+from .parser import ParsedFile, ParserRegistry, parse_file
+from .walker import iter_source_files
+
+
+@dataclass
+class FileResult:
+    parsed: ParsedFile
+    chunks: list[Chunk]
+
+
+def analyze_repo(root: str | Path) -> Iterator[FileResult]:
+    """Iterate over every supported source file under `root`, yielding the
+    parsed file + its extracted chunks.
+
+    Streaming — callers can write JSONL as it flows, without holding the
+    whole repo in memory.
+    """
+    root_path = Path(root).resolve()
+    registry = ParserRegistry()
+
+    for file_path in iter_source_files(root_path):
+        parsed = parse_file(file_path, repo_root=root_path, registry=registry)
+        if parsed is None:
+            continue
+        chunks = chunk_file(parsed)
+        yield FileResult(parsed=parsed, chunks=chunks)

karst-0.1.0/karst/ask.py

@@ -0,0 +1,149 @@
+"""Repo Q&A — question in, cited answer out.
+
+Pipeline:
+  question → embed → Qdrant top-k → assemble prompt → LLM → answer
+
+Citation discipline (spec §33): the prompt forces the model to anchor every
+claim to `file:start-end`. If no LLM is configured, callers can render the
+retrieved hits directly — still useful, just no prose synthesis.
+"""
+
+from __future__ import annotations
+
+from dataclasses import dataclass
+from pathlib import Path
+
+from .embedder import DEFAULT_MODEL, Embedder
+from .llm import LLM, LLMResponse, default_llm
+from .store import DEFAULT_COLLECTION, ChunkStore, SearchHit
+
+
+@dataclass
+class _LabeledHit:
+    """Internal: a SearchHit with a source label ('vector' or 'graph')."""
+    hit: SearchHit
+    source: str
+
+# How long any single retrieved chunk is allowed to be inside the prompt.
+# Beyond this we cut — the citation still points at the full file.
+_MAX_CHUNK_CHARS = 2_000
+
+
+@dataclass
+class AskResult:
+    question: str
+    hits: list[SearchHit]
+    answer: str | None
+    llm: LLMResponse | None
+
+
+SYSTEM_PROMPT = """\
+You are an AI Staff Engineer answering questions about a specific code repository.
+
+You are given the user's question and the top retrieved chunks of code from
+the repo. Each chunk header looks like [N] path/to/file.ts:start-end. You must:
+
+1. Answer concisely and concretely.
+2. Cite every claim with a bracketed reference in the form [path/to/file.ts:start-end].
+   Use the same path/range shown in the chunk header. Never invent files or line ranges.
+3. If the retrieved chunks do not contain enough information, say so plainly and
+   suggest what to look at next. Do not guess.
+4. Prefer evidence from the retrieved chunks over background knowledge.
+"""
+
+
+def ask(
+    question: str,
+    *,
+    storage_path: str | Path,
+    collection: str = DEFAULT_COLLECTION,
+    embedding_model: str = DEFAULT_MODEL,
+    embedder_cache_dir: str | Path | None = None,
+    top_k: int = 8,
+    llm: LLM | None = None,
+    use_llm: bool = True,
+    graph_path: str | Path | None = None,
+    graph_extra: int = 6,
+    pack_ids: list[str] | None = None,
+) -> AskResult:
+    """Question → embed → Qdrant top-k → (optional graph expansion) → LLM.
+
+    When `pack_ids` is provided, retrieval is scoped to chunks tagged with
+    any of those packs (spec §22). This is the single largest token-cost
+    lever in Phase 4 — 60-80% input reduction on big repos.
+    """
+    embedder = Embedder(
+        embedding_model,
+        cache_dir=str(embedder_cache_dir) if embedder_cache_dir else None,
+    )
+    store = ChunkStore(location=storage_path, collection=collection)
+    try:
+        (query_vec,) = embedder.embed_texts([question])
+        seed_hits = store.search(query_vec, limit=top_k, pack_ids=pack_ids)
+
+        if graph_path is not None:
+            hits = _expand_with_graph(seed_hits, graph_path, store, extra=graph_extra)
+        else:
+            hits = seed_hits
+    finally:
+        store.close()
+
+    if not use_llm:
+        return AskResult(question=question, hits=hits, answer=None, llm=None)
+
+    used_llm = llm or default_llm()
+    user_prompt = _build_user_prompt(question, hits)
+    resp = used_llm.generate(SYSTEM_PROMPT, user_prompt)
+    return AskResult(question=question, hits=hits, answer=resp.text, llm=resp)
+
+
+def _expand_with_graph(
+    seed_hits: list[SearchHit],
+    graph_path: str | Path,
+    qdrant: ChunkStore,
+    *,
+    extra: int,
+) -> list[SearchHit]:
+    """Lazy import so plain `ask` doesn't pay for networkx unnecessarily."""
+    from .graph.graphrag import expand_with_graph
+    from .graph.store import GraphStore
+
+    graph = GraphStore.load(graph_path)
+    expanded = expand_with_graph(
+        seed_hits,
+        graph=graph,
+        qdrant=qdrant,
+        max_extra=extra,
+    )
+    # Collapse back into SearchHit list so the downstream prompt builder
+    # doesn't need to learn a new type. Source label is encoded in the score
+    # rank order; graph hits will already be lower-scored than seeds.
+    return [SearchHit(chunk=h.chunk, score=h.score) for h in expanded]
+
+
+def _build_user_prompt(question: str, hits: list[SearchHit]) -> str:
+    if not hits:
+        return (
+            "No chunks were retrieved from the index for this question.\n\n"
+            f"Question: {question}\n\n"
+            "Tell the user the index is empty or the question matches nothing, "
+            "and recommend re-running `karst index <path>` or rephrasing."
+        )
+
+    parts: list[str] = ["# Retrieved chunks", ""]
+    for i, hit in enumerate(hits, start=1):
+        c = hit.chunk
+        code = c.code
+        if len(code) > _MAX_CHUNK_CHARS:
+            code = code[:_MAX_CHUNK_CHARS] + "\n… (truncated)"
+        parts.append(
+            f"[{i}] {c.citation}  "
+            f"({c.kind.value} {c.qualified_name}, score={hit.score:.3f})"
+        )
+        parts.append(f"```{c.language}")
+        parts.append(code)
+        parts.append("```")
+        parts.append("")
+    parts.append("# User question")
+    parts.append(question)
+    return "\n".join(parts)