Index of every page on this site: llms.txt
All skills
getsentry avatar

/browse-local-sentry

@a582c51 official
by Sentrygetsentry/skills1.2k stars
71
Run it yourself

Connects an agent's browser to a local Sentry dev server (dev-ui on *.dev.getsentry.net or the full devserver).

Use this Skill: https://skilld.dev/gh/getsentry/skills/browse-local-sentry

Nothing lands on disk. Nothing to clean up.

Fork this Skill

Edit a local copy. It keeps the author and licence.

SKILL.md

≈33 tokens for metadata: the name and description. ≈721 when used: this file. ≈3.4k more on demand in 3 files.

Description uses 1.7% of example budget

Before choosing a Skill, your Agent reads its name and description. All available Skills share that space.

  • A shorter description leaves more room for other Skills. This entry exceeds our 1% size suggestion.

In our Claude Code example, all Skill names and descriptions share 8,000 characters. This Skill uses ≈133 characters, or 1.7%.

The 1% threshold is a size suggestion. Longer descriptions can still fit.

Your model, settings, and other Skills decide how much text your Agent can read.

Example settings and source

The example uses a 200k-token context and default Claude Code settings. The count includes the name, description, separators, and when_to_use when present. Codex also counts local file paths.

Skit's source and limits: Codex 0.160.1, Claude Code 2.1.292.

Browse Local Sentry

Server URL Data Browser
dev-ui (pnpm run dev-ui) https://<org-slug>.dev.getsentry.net:<port>/. Ask the user for the org if it is unknown. Production sentry.io, as the user The user's own Chrome, the only one with their session
Full devserver (devservices serve) http://dev.getsentry.net:8000/ Local database Any

Run every check in every session. Chrome updates, re-auth, and port changes undo earlier setup without warning. When a check fails, open the named section of references/troubleshooting.md.

Rules

  • dev-ui acts on production sentry.io as the user, including any staff or superuser session. Do not create, edit, or delete anything, and stay in the org the user named, unless the user approves the specific action. Use the full devserver for flows that change data.
  • Never bypass certificate warnings (thisisunsafe, --ignore-certificate-errors, --acceptInsecureCerts), log in to sentry.io yourself, or type the user's credentials.
  • Change the user's MCP config only with their approval.
  • In the user's own Chrome, open your own tabs with background: true, never isolatedContext, which has no session. Never navigate, reload, close, read, screenshot, run scripts in, or quote the user's tabs. Emulate viewports instead of resizing the window, and close your tabs when done.

Step 1: Find the Server

Use the URL the user gives. Otherwise never assume a port: dev-ui tries 10 ports upward from SENTRY_WEBPACK_PROXY_PORT (default 7999) and prints the one it took, and the devserver logs to .artifacts/dev.log. List listeners with lsof -nP -iTCP -sTCP:LISTEN and match one to its clone with lsof -a -p <pid> -d cwd.

Step 2: Check the Certificate (dev-ui Only)

Outside any shell sandbox, run curl -sS -o /dev/null https://dev.getsentry.net:<port>/ (on macOS, /usr/bin/curl, which reads the keychain). If it fails, see Certificate.

Step 3: Attach to the User's Chrome (dev-ui Only)

Call list_pages, or use another tool that already drives the user's signed-in Chrome. The browser is attached when it lists the user's own tabs. If not, see Chrome.

Step 4: Check Authentication

Open the server URL in a new tab and read /api/client-config/. Continue when isAuthenticated is true. If not, see Authentication.

After editing frontend source, wait for the rebuild, then reload with the cache bypassed. If the console shows "An SSL certificate error occurred when fetching the script.", the certificate is not trusted; return to Step 2.

When a Step Is Blocked

Stop and tell the user which step failed and the exact action they need to take. Do not substitute tests, stories, or reasoning for browser QA, and do not write "browser verification was blocked" into a PR description in place of doing it.

Source: SKILL.md on GitHub

skilld matched fixed text patterns in SKILL.md and file names. Patterns miss obfuscated code.

skilld run checks every file with the same patterns. It asks for approval before it loads a Skill with a behavior marked Needs approval.

No alerts15d3 checks · Risk SAFE
  • Gen Agent Trust Hub15d

    This skill provides a safe and well-documented workflow for Sentry developers to connect an agent to their local development environment. It uses standard diagnostic tools and version-pinned browser automation via an MCP server. Security is maintained through explicit safety rules, constraints on production data access, and requirements for user approval before making persistent configuration changes.

  • Socket15d

    No alerts

  • Snyk15d

    Risk: LOW · No issues

Signed by skilld at a582c51. This ties the file your Agent reads to that commit on GitHub. It does not review the instructions.

Last checked against GitHub 6 days ago.

Activeupdated 4 weeks ago

README badge

README badge for getsentry/skills/browse-local-sentry