{ "SchemaVersion": "1", "Kind": "DirectoryIssues", "Slug": "ivygrep", "Name": "ivygrep", "CanonicalUrl": "https://askpod.ai/mcp/ivygrep/issues", "ServerUrl": "https://askpod.ai/mcp/ivygrep", "IssueTotal": 8, "Held": 8, "Issues": [ { "Title": "[Bug]: regex search ignores type and context options", "Excerpt": "## What happened?\n\nCLI, MCP, and Web accept type/context options for regex search, but `DaemonRequest::RegexSearch` and local regex implementation carry neither field.\n\n## Minimal reproduction\n\n```shell\nig --regex --type markdown --context 10 --json --limit 1 \\\n 'pub fn regex_search' .\n```\n\nObserved Rust file `src/regex_search.rs` with `start_line == end_line`. Both Markdown filter and context 10 were ignored.\n\n## Expected behavior\n\nRegex mode honors same documented type filter and context cont", "SourceUrl": "https://github.com/bvolpato/ivygrep/issues/214", "PublishedAt": "2026-08-04T05:33:42.000Z", "State": "closed", "Comments": 0, "Reporter": "Maintainer", "Rank": "top", "Extractor": "github_issue" }, { "Title": "MCP: keep coding-agent indexes fresh and simplify setup", "Excerpt": "## Problem\n\nAn MCP-only coding-agent session can auto-index a workspace but does not keep a watcher alive. Edits made after the first query can therefore remain absent until another CLI operation reconciles the index. On Windows, daemon auto-spawn also rejects the packaged `ig.exe` filename.\n\nThe agent setup documentation has drifted across clients, and release installation requires too many manual steps.\n\n## Reproduction\n\n1. Start `ig --mcp` with an isolated `IVYGREP_HOME`.\n2. Call `ig_search` ", "SourceUrl": "https://github.com/bvolpato/ivygrep/issues/135", "PublishedAt": "2026-06-17T20:15:52.000Z", "State": "closed", "Comments": 0, "Reporter": "Maintainer", "Rank": "top", "Extractor": "github_issue" }, { "Title": "P1: MCP server reloads the neural embedding model on every request", "Excerpt": "## Performance (P1)\n\n`src/mcp.rs::execute_ivygrep_search` calls `create_model(false)` on **every** `ig_search` request. `serve_stdio` is a long-lived process, so this reconstructs the Candle/ONNX model (load tokenizer + weights) per call — hundreds of ms of avoidable latency on every search, plus memory churn.\n\nThe daemon already loads the model once into an `Arc>` (`DaemonState.lazy_model`) and `cached_hash_model()` uses the same pattern. The MCP server should cache its query model", "SourceUrl": "https://github.com/bvolpato/ivygrep/issues/57", "PublishedAt": "2026-05-25T04:26:58.000Z", "State": "closed", "Comments": 0, "Reporter": "Maintainer", "Rank": "top", "Extractor": "github_issue" }, { "Title": "P0: MCP auto-index embeds inline with the neural model → first query hangs on large repos; parallel MCP melts the host", "Excerpt": "## Stability / Performance (P0)\n\nThe MCP auto-index path embeds **every chunk with the neural ONNX model inline**, on the first query. `src/mcp.rs::execute_ivygrep_search` does:\n\n```rust\nlet model = create_model(false); // neural (384-dim), loaded per request\n...\nif !workspace_is_indexed(¤t_workspace) {\n let _summary = index_workspace(¤t_workspace, model.as_ref())?; // <-- neural model passed to the indexer\n}\n```\n\n`index_workspace` then calls `embedding_model.embed_ba", "SourceUrl": "https://github.com/bvolpato/ivygrep/issues/56", "PublishedAt": "2026-05-25T04:16:27.000Z", "State": "closed", "Comments": 0, "Reporter": "Maintainer", "Rank": "top", "Extractor": "github_issue" }, { "Title": "A search panic crashes the entire MCP server (no panic isolation)", "Excerpt": "## Robustness (P1)\nThe MCP server runs single-threaded with no panic isolation (`src/mcp.rs`): a panic anywhere in `dispatch`→`execute_ivygrep_search`→search unwinds through `serve_stdio` and **crashes the whole MCP session**. (The daemon isolates each connection in a tokio task; MCP does not.) Wrap tool-call handling in `catch_unwind` and return a JSON-RPC error.", "SourceUrl": "https://github.com/bvolpato/ivygrep/issues/53", "PublishedAt": "2026-05-23T17:06:32.000Z", "State": "closed", "Comments": 0, "Reporter": "Maintainer", "Rank": "top", "Extractor": "github_issue" }, { "Title": "Unbounded request/Content-Length sizes on MCP and daemon transports (OOM DoS)", "Excerpt": "## Robustness / DoS (P1)\n- MCP `Content-Length` framing allocates `vec![0u8; len]` with `len` straight from the client header and **no upper bound** (`src/mcp.rs`) → trivial OOM.\n- Both the daemon (`src/daemon.rs`) and MCP use unbounded `read_line`, so a client streaming a huge line without a newline grows memory without bound.\n\nCap request/frame size and reject oversized input with an error instead of allocating.", "SourceUrl": "https://github.com/bvolpato/ivygrep/issues/51", "PublishedAt": "2026-05-23T17:06:31.000Z", "State": "closed", "Comments": 0, "Reporter": "Maintainer", "Rank": "top", "Extractor": "github_issue" }, { "Title": "MCP --all search returns no results for indexed workspaces (CLI --all works)", "Excerpt": "## Finding (unfixed — needs investigation)\nThe MCP server's `--all` (all-indices) search returns **zero results** even when the target workspaces are indexed and registered (verified: `ig --status` lists them with `last_indexed_at_unix` set), whereas the **CLI** `ig --all` finds the same content correctly with absolute paths.\n\nReproduction:\n1. `ig --add repoA --hash --no-watch`; `ig --add repoB --hash --no-watch` (same `IVYGREP_HOME`).\n2. CLI: from repoA, `ig --all --literal \"