project-lsp
vv3by nexaedge · author: Jaison Erick
Go, Python, Rust and TypeScript code intelligence for Claude Code (go to definition, find references, symbols, hover, diagnostics) that runs one language server per project, so files of other repositories and git worktrees resolve their own imports. Each server starts in its project through mise, with that project's tool versions and virtualenv, and TypeScript 7 projects get the server TypeScript 7 ships.
What project-lsp does, in the publisher's words
Code intelligence for Claude Code in Go, Python, Rust, TypeScript and JavaScript (go to definition, find references, document and workspace symbols, hover, call hierarchy, diagnostics after each edit) that stays right when a session works on files of several repositories or git worktrees.
Claude Code starts one language server per language for the whole session, rooted at the folder the session started in. A file of another repository, or of a worktree under.claude/worktrees/, then reads as outside that server's project: its imports fail to resolve and every edit comes back with diagnostics that are not true. project-lsp stands in front of the real server and runs one per project instead.
- The project of a file is the git worktree that holds it, the folder git rev-parse --show-toplevel names from the file's directory. A file outside git belongs to its own directory.
From the project README, punctuation lightly normalized · Open the README on GitHub
- Skills
- 0
- MCP servers
- 0
- Stars
- 1
- License
- n/a
- Repo created
- 2026-03-05
- Last pushed
- 2026-10-01
- Publisher type
- Organization
- Version
- v3
Links
Category
Other. Assigned by this directory. Agent Plugins 1.0.0 has no category field, so no manifest declares one.
Keywords
lsp · language-server · go · python · rust · typescript · worktree
Related by publisher and keywords
Publish a self-contained HTML file as a shareable artifact on artifacts.nexaedge.com. Sign in with any Google Workspace account: the artifact belongs to that account's domain, and only that domain can open it. Also issues expiring links for people outside the domain.
Manage Cloudflare DNS zones and records through Terraform in the nexaedge/infrastructure repository. Automates the full workflow: edit Terraform files, open PR, review plan, merge, watch apply, and verify.
Turn finished analysis results into a self-contained interactive HTML report: charts, cohort heatmaps, sortable tables and stat tiles, all inline in one file that opens by double-click with no network. Carries no brand of its own: it derives a theme from a brand color, or lifts one verbatim out of a report it built earlier.
Spec-driven project execution pipeline backed by Linear. Same skills and agents as spec-plugin, but writes specs, stories and validation to Linear via GraphQL while keeping reasoning artifacts on disk.
Read any Miro board you can open in your browser, then work out what it says. Miro's public API cannot reach a board you were invited to in someone else's team, and the canvas is WebAssembly so there is nothing in the page to scrape; this reads the board through your own signed-in session instead, opening Chrome once and then talking to Miro's realtime gateway directly. Extraction and interpretation are kept apart on purpose: the scripts measure geometry, colour, containment and edge alignment and never conclude anything, offering overlapping candidate groupings with the evidence behind each, because what a board means lives in how it was arranged and only reading the text can settle it.
Front end for the nexaedge marketplace that works the same on Claude Code, Codex and Antigravity. Detects which agent is running, translates to that agent's native install path, and keeps everything at user scope so a plugin is installed once per machine rather than copied into each repository.
[](https://agentpluginsdirectory.com/plugins/project-lsp)✓ Verified . We fetched the manifest from GitHub and checked it against the official Agent Plugins 1.0.0 schema at agent-plugins.org.