getsentry-sentry-skills

Sentry Skills

Note For skills to help set up Sentry in your project or debug production issues, see https://github.com/getsentry/sentry-for-ai

Agent skills for Sentry employees, following the Agent Skills open format.

Installation

Claude Code

claude plugin marketplace add getsentry/skills
claude plugin install sentry-skills@sentry-skills

If you use claude plugin marketplace add --sparse for this repo, include skills and agents alongside .claude-plugin because the root plugin manifest loads repo-root skills/ and agents/.

Restart Claude Code after installation. Skills activate automatically when relevant.

Update:

claude plugin marketplace update
claude plugin update sentry-skills@sentry-skills

Or run /plugin to open the plugin manager.

Skills Package (skills.sh)

For agents supporting the skills.sh ecosystem:

npx skills add getsentry/skills

Works with Claude Code, Cursor, Cline, GitHub Copilot, and other compatible agents.

Available Skills

Skill	Description
agents-md	Create and maintain concise, reference-backed AGENTS.md and CLAUDE.md files.
blog-writing-guide	Write, review, and improve blog posts for the Sentry engineering blog following Sentry's specific writing standards, voice, and quality bar.
brand-guidelines	Write copy following Sentry brand guidelines.
claude-settings-audit	Analyze a repository to generate recommended Claude Code settings.json permissions.
code-review	Perform code reviews following Sentry engineering practices.
code-simplifier	Simplifies and refines code for clarity, consistency, and maintainability while preserving all functionality.
commit	ALWAYS use this skill when committing code changes — never commit directly without it.
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 users through a structured workflow for co-authoring documentation.
document-api-endpoint	Document and type a Sentry API endpoint with drf-spectacular OpenAPI schema — write/fix @extend_schema, specify response TypedDicts, type parameters, fix type drift, and promote to PUBLIC.
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 specified team or opened by a team member.
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 presentation slides using 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, lay them out as before/after tables in PR descriptions, and upload them with gh --attach.
pr-writer	Create, refresh, and rewrite pull request titles and descriptions following Sentry conventions.
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 such as dataclass repr, asdict, model_dump, JSON.stringify, logs, and tracing spans.
security-model	Define and maintain SECURITY_MODEL.md files that tell security review tools a repository's trust boundaries, invariants, reportable finding classes, and exclusions.
security-review	Security code review for vulnerabilities.
skill-scanner	Scan agent skills for security issues.
skill-writer	Canonical workflow to synthesize, create, and iteratively improve agent skills for this repository.
sred-project-organizer	Take a list of projects and their related documentation, and organize them into the SRED format for submission.
sred-work-summary	Go back through the previous year of work and create a Notion doc that groups relevant links into projects that can then be documented as SRED projects.
triage-frontend-issues	Triage the Sentry javascript project queue by archiving non-actionable noise (third-party libs, browser quirks, transient 5xx, test traffic) with untilEscalating after user approval.
typing-exclusion-worker	Python typing exclusion worker: remove assigned mypy exclusion modules in small scoped batches, fix typing issues, run validation, and produce a structured completion summary.

Available Subagents

Subagent	Description
code-simplifier	Simplifies and refines code for clarity, consistency, and maintainability while preserving all functionality
senpai	Senior engineer and technical mentor for new Sentry hires. Explains infrastructure, architecture, and engineering concepts step-by-step with references

Contributing

Local Development

git clone git@github.com:getsentry/skills.git ~/sentry-skills
claude plugin marketplace add ~/sentry-skills
claude plugin install sentry-skills

If you use claude plugin marketplace add --sparse for this repo, include skills and agents alongside .claude-plugin because the root plugin manifest loads repo-root skills/ and agents/.

Repository Structure

sentry-skills/
├── .claude-plugin/
│   ├── marketplace.json      # Marketplace manifest
│   └── plugin.json           # Root plugin manifest for sentry-skills
├── .agents/
│   └── skills -> ../skills   # Local mirror for agent tooling
├── agents/
│   ├── code-simplifier.md
│   └── senpai.md
├── skills/
│   ├── code-review/
│   │   └── SKILL.md
│   └── commit/
│       └── SKILL.md
├── AGENTS.md                 # Agent-facing documentation
├── CLAUDE.md                 # Symlink to AGENTS.md
└── README.md                 # This file

The canonical skill source for the sentry-skills plugin lives at repo-root skills/. .agents/skills mirrors that tree for local agent tooling, and the root plugin manifest loads repo-root skills/ and agents/.

New skills and materially changed skills should include a root-level SPEC.md alongside SKILL.md.

Creating New Skills

Skills follow the Agent Skills specification. Each skill requires a SKILL.md file with YAML frontmatter.

For repeatable skill-writer evaluation prompts, see skills/skill-writer/EVAL.md.

Skill Template

Create a new directory under skills/:

skills/my-skill/
├── SKILL.md
└── SPEC.md

SKILL.md format:

---
name: my-skill
description: A clear description of what this skill does and when to use it. Include keywords that help agents identify when this skill is relevant.
---

# My Skill Name

## Instructions

Step-by-step guidance for the agent.

## Examples

Concrete examples showing expected input/output.

## Guidelines

- Specific rules to follow
- Edge cases to handle

SPEC.md format:

Use SPEC.md as the maintenance contract for the skill. It should describe intent, scope, trigger context, source/evidence model, reference architecture, evaluation expectations, known limitations, and maintenance notes. Keep runtime instructions in SKILL.md; keep source inventories in SOURCES.md; keep durable examples in references/evidence/.

Naming Conventions

• name: 1-64 characters, lowercase alphanumeric with hyphens only
• description: Up to 1024 characters, include trigger keywords
• Keep SKILL.md under 500 lines; split longer content into reference files

Optional Fields

Field	Description
license	License name or path to license file
compatibility	Environment requirements (max 500 chars)
allowed-tools	Space-separated list of tools the skill can use. Do not use commas: Warden drops comma-suffixed names like Read,
metadata	Arbitrary key-value pairs for additional properties

---
name: my-skill
description: What this skill does
license: Apache-2.0
allowed-tools: Read Grep Glob
---

Where Skills Belong

Skills should live in the appropriate location based on their scope:

Scope	Location	Example
Global - Used across Sentry	sentry-skills plugin	commit, code-review, pr-writer
Domain-specific - Used by a team or domain	Dedicated plugin in this repo (e.g., infra-skills)	gcp-logs, terraform-review
Repo-specific - Only relevant to one repo	The repository itself (.claude/skills/)	Project-specific workflows

When deciding where to place a skill:

• If most Sentry engineers would benefit, add it to sentry-skills
• If only a specific team needs it, create or use a domain-specific plugin
• If it only makes sense in one repo, keep it in that repo

Marketplace Structure

This repository is a Claude Code marketplace - a collection of plugins that can be installed independently. The marketplace manifest (.claude-plugin/marketplace.json) lists all available plugins:

{
  "plugins": [
    {
      "name": "sentry-skills",
      "source": "./"
    },
    { "name": "infra-skills", "source": "./plugins/infra-skills" }
  ]
}

Marketplace entries point at a plugin source directory. That source can be a dedicated plugin directory with its own .claude-plugin/plugin.json, or repo root when the root also carries a plugin manifest. The sentry-skills entry in this repo uses repo root as its plugin source so it can keep canonical skills/ and agents/ at top level.

Users can install individual plugins:

# Install just the global skills
claude plugin install sentry-skills@sentry-skills

# Install domain-specific skills
claude plugin install infra-skills@sentry-skills

To add a new domain-specific plugin:

1. Create plugins/<plugin-name>/.claude-plugin/plugin.json
2. Add skills under plugins/<plugin-name>/skills/
3. Register the plugin in .claude-plugin/marketplace.json

Vendoring Skills

We vendor (copy) skills and agents that we use regularly into this repository rather than depending on external sources at runtime. This approach:

• Ensures consistency - Everyone on the team uses the same version of each skill
• Enables customization - We can adapt skills to Sentry-specific conventions
• Improves reliability - No external dependencies that could change or disappear

Attribution

When vendoring a skill or agent from an external source, retain proper attribution: