agentleFS
jamesyang124jamesyang124/agent-skills/skills/install-sentry-mcp/SKILL.md

install-sentry-mcp

Skill2 starsChanged 13 days agoMITView on GitHub

Install and configure the official Sentry MCP server for Claude or GitHub Copilot agents. Supports remote (OAuth, recommended) and local stdio (npx + access token) methods. Use when setting up Sentry MCP, error-monitoring MCP, Seer, or asked to install sentry mcp.

  • Reads credentials
  • Installs packages

The file

We split this skill into the rules it gives an agent

4 rules. About 0 are common, 2 are distinctive, and 2 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈2

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 2

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Prefer remote unless the user is on self-hosted Sentry or explicitly needs the local process.

    line 28
  2. Distinctive:

    All optional — anything not passed is asked in Step 1.

    line 32
  3. Very specific:

    Create a user auth token at: Sentry → Settings → Auth Tokens (https://sentry.io/settings/account/api/auth-tokens/).

    line 93
  4. Very specific:

    Run setup-project-config after installing to record them.

    line 147

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Install Sentry MCP

Install this skill

npx skills add jamesyang124/agent-skills --skill install-sentry-mcp -g

Adds Sentry access to your agent: list/inspect issues and events, read stack traces and breadcrumbs, query traces, and trigger Seer AI root-cause analysis. Official repository: https://github.com/getsentry/sentry-mcp — hosted endpoint: https://mcp.sentry.dev/mcp

Two methods

MethodAuthSecret storedWhen
remote (recommended)browser OAuth to mcp.sentry.devnoneSaaS sentry.io; simplest, most tools, auto-updated
stdiouser auth token~/.env.mcp-sentryself-hosted Sentry, or when a local process is required

Prefer remote unless the user is on self-hosted Sentry or explicitly needs the local process.

Arguments

All optional — anything not passed is asked in Step 1.

TokenMeaning
claude / copilotwhich agent to configure
remote / stdioinstall method (remote OAuth recommended; stdio for self-hosted Sentry)

The access token is never accepted as an argument. For Copilot stdio, VS Code asks for it in a password prompt when the server starts. For Claude stdio, the installer collects it in the terminal.


Quick Start

Collect parameters, then run the bundled script. Never request or pass an access token through a chat or command-line argument.

Step 1 — Collect parameters

Use vscode_askQuestions with:

  1. agent — which agent to configure
    • options: claude, copilot
  2. method — installation method
    • options: remote (recommended, OAuth, no token), stdio (local, needs token)
  3. host — (stdio + self-hosted only) Sentry hostname, e.g. sentry.example.com. Omit for SaaS sentry.io.
  4. scope — project (default) or user (Copilot only). Use user to register Sentry globally for all VS Code workspaces.
  5. disabled skills — (optional) comma-separated Sentry MCP skills to disable, e.g. seer.

Step 2 — Run the installer

PROJECT_ROOT="<workspace-root>" \
  bash "<skill-dir>/scripts/install.sh" \
  --agent <agent> \
  --method <method>

<skill-dir> is this skill's folder, the base directory shown when the skill loads (for example ~/.claude/skills/install-sentry-mcp).

For a self-hosted Sentry MCP available across all VS Code workspaces:

bash ~/.copilot/skills/install-sentry-mcp/scripts/install.sh \
  --agent copilot \
  --method stdio \
  --host sentry.example.com \
  --scope user \
  --disable-skills seer
  • remote: writes only the MCP registration (a url entry). No credentials stored — the agent runs a browser OAuth flow to https://mcp.sentry.dev/mcp on first use.
  • Copilot stdio: writes a password-protected VS Code input reference. The token is entered only when VS Code starts the server and is never written to the MCP configuration file.
  • Claude stdio: prompts for a Sentry user auth token (hidden input), stores it globally in ~/.env.mcp-sentry (chmod 600), and launches the MCP process through that protected file. Add --host sentry.example.com for self-hosted.
  • The script merges into any existing config file; it will not overwrite other servers.

Create a user auth token at: Sentry → Settings → Auth Tokens (https://sentry.io/settings/account/api/auth-tokens/). Scopes: org:read, project:read, project:write, team:read, team:write, event:write.

Step 3 — Confirm

Tell the user: Restart your agent to load Sentry MCP. For the remote method, the first Sentry tool call opens a browser to authorize.

Prerequisites

  • remote: none beyond an agent that supports remote/HTTP MCP servers.
  • stdio: Node.js ≥ 18 (node -v) so npx @sentry/mcp-server@latest can run. jq for config merge (brew install jq).

What Gets Written

Remote — Claude (.mcp.json) uses mcpServers; Copilot (.vscode/mcp.json) uses servers:

{
  "mcpServers": {
    "sentry": { "type": "http", "url": "https://mcp.sentry.dev/mcp" }
  }
}

Copilot stdio — resolves npx to its absolute path so VS Code can launch it even when its GUI environment has not loaded NVM. The token is requested as a password input:

{
  "servers": {
    "sentry": {
      "command": "/absolute/path/to/npx",
      "args": ["-y", "@sentry/mcp-server@latest"],
      "env": {
        "SENTRY_ACCESS_TOKEN": "${input:sentry-access-token}",
        "SENTRY_HOST": "sentry.example.com"
      }
    }
  },
  "inputs": [{
    "id": "sentry-access-token",
    "type": "promptString",
    "description": "Sentry access token",
    "password": true
  }]
}

SENTRY_HOST is omitted for SaaS sentry.io.

Per-project settings

Auth and connection are global (OAuth / ~/.env.mcp-sentry). Which Sentry org/project this repo maps to is per-project — captured by /setup-project-config in docs/agents/sentry.md (Org Slug, Project Slug). Run setup-project-config after installing to record them.

Troubleshooting

  • node/npx not found → install Node.js ≥ 18, then restart your shell (stdio only)
  • jq not found → brew install jq on macOS
  • OAuth window does not open (remote) → confirm the agent supports remote MCP; re-trigger a Sentry tool call
  • 401/auth errors (stdio) → regenerate the token and re-run; verify the scopes listed above
  • Self-hosted not reachable → confirm --host is the hostname only (no https://, no path)
  • Copilot token prompt does not appear → reload VS Code, then start the sentry server from MCP server management. Enter the token in the password prompt, not in mcp.json or chat.
  • Config not loading → confirm the agent was fully restarted after install

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.