plugins-claude/serena/skills/serena-setup/SKILL.md
Install the Serena MCP server. Use on requests to install/setup Serena, or when `mcp__serena__*` tools are unavailable. Installs via `uv tool install` (Python) and registers in `~/.claude.json`.
npx skillsauth add st0nefish/claude-toolkit serena-setupInstall this skill globally with one command. Works with Claude Code, Cursor, and Windsurf.
3 of 9 scanners reported clean
Some scanners were skipped, did not run, or reported a non-clean status. Review each row below.
Serena is a Python project distributed via the oraios/serena GitHub repo.
It exposes a stdio MCP server (serena start-mcp-server) that brings
language-server-driven code intelligence into Claude Code.
uv (Astral's Python package manager) — install via:
curl -LsSf https://astral.sh/uv/install.sh | sh
Or via system package manager (Arch: pacman -S uv).
Serena's source tree includes Node.js artifacts only as a transitive concern; this skill never touches npm/node/npx. Installation is pure Python via uv.
uv tool install --from git+https://github.com/oraios/serena serena
This puts the serena binary at ~/.local/bin/serena (ensure that's on
PATH). Verify:
which serena && serena --version
To upgrade later: uv tool upgrade serena.
Add an MCP server entry to ~/.claude.json. The recommended invocation:
{
"mcpServers": {
"serena": {
"type": "stdio",
"command": "serena",
"args": [
"start-mcp-server",
"--context",
"claude-code",
"--project-from-cwd"
],
"env": {}
}
}
}
Key flags:
--context claude-code — Serena tunes its behavior (prompts, tool descriptions) for Claude Code.--project-from-cwd — Serena auto-detects the active project from the current
working directory. No per-project config needed.The entry can live at user level (~/.claude.json top-level mcpServers) for
all projects, or per-project under .projects[<path>].mcpServers.
Serena's "memories" feature stores opaque sidecar notes that aren't source-controlled. Disable it for source-of-truth-in-the-repo workflows.
Add to your Serena project config (.serena/project.yml or via CLI flag):
# .serena/project.yml
disable_memories: true
disable_onboarding: true
Or pass --disable-memories --disable-onboarding to the start-mcp-server
args. With memories disabled, the memory MCP tools (write_memory,
read_memory, etc.) won't appear in the tool list, and the onboarding flow
won't trigger.
After registering and reconnecting Claude Code:
mcp__serena__* tools should appear in the available tool list (some
surfaces show them as "deferred" until first use).mcp__serena__initial_instructions should return Serena's
manual without errors.mcp__serena__get_symbols_overview on a source file in your
project should return structured symbol data.If tools don't appear, check:
serena is on PATH (which serena)"type": "stdio" and the correct args~/.serena/logs/ by defaultserena-cheatsheet skill (this plugin) covers usage quirks once Serena
is running.serena-explorer agent (this plugin) handles verbose meta-analysis
queries.development
Start work from your description — explore the codebase and plan
data-ai
Multi-phase, multi-agent feature workflow: spec → plan → refine → divide → execute → review. Invoke when the user escalates a session-start/session-issue flow to orchestration, or asks to run a non-trivial feature (multiple files, design ambiguity, cross-cutting concerns, correctness-critical paths) through the full multi-agent workflow. For small fixes, prefer session-start.
tools
Browse open issues, pick one, and start work on it
tools
Interact with GitHub and Gitea issue trackers and CI systems. List and show issues, file bugs, comment on issues or PRs, list and show pull requests, and fetch CI run logs — all from any repo context without leaving the session.