Sentry Skills

# Sentry Skills: setup notes

This kit holds setup notes written by Mindket. It does not contain the skills, their scripts or the two subagents. Those install from Sentry’s repository: https://github.com/getsentry/skills

License: Apache-2.0. Full text: https://github.com/getsentry/skills/blob/main/LICENSE. Two skills, security-review and django-access-review, carry reference material based on the OWASP Cheat Sheet Series (CC BY-SA 4.0), named in their SKILL.md files.

Sentry Skills is made by Sentry. Mindket lists it and is not Sentry.

The pages these notes came from: the README (https://github.com/getsentry/skills/blob/main/README.md), each skill’s SKILL.md under https://github.com/getsentry/skills/tree/main/skills, AGENTS.md, and the plugin files in https://github.com/getsentry/skills/tree/main/.claude-plugin. Facts about Claude Code’s plugins come from its own docs, named where they are used. Mindket copied every command and ran none of them.

## What it is

Thirty-one skills and two subagents that Sentry’s engineers use with coding agents. The README calls them “Agent skills for Sentry employees”. They follow the open Agent Skills format. A skill is a folder with a SKILL.md file of instructions that the agent reads when your request matches.

The README points elsewhere for the other kind of Sentry help: for skills that set up Sentry in your project or debug production issues, see https://github.com/getsentry/sentry-for-ai

## The skills

| Skill | What it does (the README’s words, shortened, and for commit its SKILL.md’s) |
|---|---|
| agents-md | Create and maintain concise, reference-backed AGENTS.md and CLAUDE.md files |
| blog-writing-guide | Write and review posts for the Sentry engineering blog |
| brand-guidelines | Write copy following Sentry’s brand guidelines |
| browse-local-sentry | Connect an agent’s browser to a local Sentry dev server, to QA and screenshot the real product |
| claude-settings-audit | Analyze a repository to generate recommended Claude Code settings.json permissions |
| code-review | Review code following Sentry engineering practices |
| code-simplifier | Simplify and refine code for clarity while preserving all functionality |
| commit | Create Sentry-style conventional commits |
| create-branch | Create git branches following Sentry naming conventions |
| django-access-review | Django access control and IDOR security review |
| django-perf-review | Django performance code review |
| doc-coauthoring | Guide a structured workflow for co-authoring documentation |
| document-api-endpoint | Document and type a Sentry API endpoint with a drf-spectacular schema |
| find-bugs | Find bugs, security vulnerabilities and code quality issues in local branch changes |
| gh-review-requests | Fetch unread GitHub notifications for open PRs where review is requested from a team |
| gha-security-review | GitHub Actions security review for workflow exploitation vulnerabilities |
| iterate-pr | Iterate on a PR until CI passes and actionable review feedback is addressed |
| presentation-creator | Create data-driven slides with React, Vite and Recharts, with Sentry branding |
| pr-link-issue | Append a GitHub issue link and its Linear ticket to the current PR’s description |
| pr-screenshots | Capture real-product UI screenshots for before and after tables in PR descriptions |
| pr-writer | Create, refresh and rewrite pull request titles and descriptions |
| prompt-optimizer | Optimize prompts with evals, model-family adapters and exact external context references |
| secret-serialization | Find secrets and tokens that leak through generated serialization, logs and tracing spans |
| security-model | Define and maintain SECURITY_MODEL.md files for security review tools |
| security-review | Security code review for vulnerabilities |
| skill-scanner | Scan agent skills for security issues |
| skill-writer | Create and improve agent skills for the repository |
| sred-project-organizer | Organize a list of projects into the SRED format for submission |
| sred-work-summary | Group the previous year of work into projects as a Notion doc |
| triage-frontend-issues | Triage the Sentry javascript project queue by archiving non-actionable noise, after approval |
| typing-exclusion-worker | Remove assigned mypy exclusion modules in small batches |

The two subagents are code-simplifier and senpai, a senior engineer and mentor for new Sentry hires.

Source: the README’s tables “Available Skills” and “Available Subagents”.

## Which skills fit outside Sentry

Mindket’s reading of the SKILL.md files, not the README’s.

- Written for anyone: security-review, find-bugs, django-perf-review, django-access-review, gha-security-review, secret-serialization, skill-scanner, iterate-pr, agents-md, claude-settings-audit, code-simplifier, doc-coauthoring, prompt-optimizer, typing-exclusion-worker (it works one batch of a typing clean-up that someone hands it).
- Sentry’s own conventions, useful as a model or when you work to the same rules: commit, create-branch, pr-writer, pr-screenshots (its description names the Sentry UI), code-review, skill-writer (the README calls it the workflow for this repository’s skills).
- Tied to Sentry’s staff, systems or code, so of little use elsewhere: blog-writing-guide, brand-guidelines, browse-local-sentry, document-api-endpoint, gh-review-requests, presentation-creator, pr-link-issue, sred-project-organizer, sred-work-summary, triage-frontend-issues.

## Files in this kit

- setup/claude-code.md: the plugin route, the update commands and the contributor route
- setup/other-agents.md: the npx skills command for Cursor, Cline, GitHub Copilot and other agents
- setup/first-run.md: what the skills need (gh, uv, MCP servers) and what each one changes

## What you need first

- A coding agent. The README names Claude Code, and says the npx route works with Claude Code, Cursor, Cline, GitHub Copilot and other compatible agents.
- Node.js, for the npx route, and for presentation-creator, whose SKILL.md runs npm.
- The GitHub command line tool, gh, signed in, for pr-writer, iterate-pr, gh-review-requests, pr-link-issue and pr-screenshots, and for find-bugs when it reads the default branch.
- uv, a Python package manager, for the skills that run Python scripts: iterate-pr, skill-scanner, gh-review-requests and security-model.
- Notion, Linear and Sentry MCP servers, only for the Sentry-staff skills that name them (setup/first-run.md).
- No account, sign-up or API key for the pack itself.

## What installs where

- Plugin route: Claude Code keeps the plugin’s files itself. One plugin, sentry-skills, holds all the skills and both subagents, so you get them together.
- npx skills route: the skill folders in your agent’s skills folder.

## Your data

A skill is text and sends nothing by itself. The agent runs the commands a skill names, so what leaves your machine is what those commands send: gh talks to GitHub, and the MCP skills talk to Sentry, Notion or Linear. Some skills write: commit and iterate-pr (it commits, pushes and can reply to review threads) act on your repository. pr-writer opens a draft pull request on GitHub and updates an existing one’s title and body, and pr-link-issue edits a PR’s body. sred-work-summary writes a private Notion document.

## Update and remove

setup/claude-code.md gives the README’s update commands. The README gives no removal step. Claude Code’s docs say what to do for a plugin, named in setup/claude-code.md.

README.md  7 KB
setup/claude-code.md  3 KB
setup/first-run.md  4 KB
setup/other-agents.md  2 KB