Warden Skills
Security review skills for Sentry Warden: authorization flaws, code execution, data leaks, DoS, GitHub Actions and PII.
- 58 ★ on the repository
- Last commit
- MIT
Minimal access
Why this level
- The skills describe searching for vulnerabilities in code and make no changes themselves
Install
- Works in
In your terminal, with SkillFoxx CLI
npx skillfoxx add skills/warden-skillsDetects the agents on your machine, shows the access level and pins the version.
Other ways to install
Access and checks
How we reviewMinimal accessThe tool runs nothing and makes no network calls.
- The skills describe searching for vulnerabilities in code and make no changes themselves
This is third-party code. Review the repository files before installing.
What it does
A set of generalized skills for Warden, a Sentry tool that reviews code changes. Each skill states in `SKILL.md` what to look for, and Warden runs it against your code. The set has seven skills. `wrdn-authz` finds authorization flaws: IDOR, missing owner and tenant scoping, role checks that fail open, privilege escalation and mass assignment. `wrdn-code-execution` catches paths from untrusted input to code and command execution on the server. `wrdn-data-exfil` finds SSRF, path traversal, SQL injection and other leaks. `wrdn-dos-review` looks for denial of service, `wrdn-gha-workflows` for GitHub Actions workflow vulnerabilities and `wrdn-pii` for personal and customer data in changes. `vercel-deepsec` gives a broad web application vulnerability scan.
Who it is for. Teams that connected Warden to pull requests and want ready vulnerability rules instead of writing their own.
Good fit when
- A pull request adds an endpoint and you must check whether it exposes other users’ data by ID
- A GitHub Actions workflow uses `pull_request_target` and needs an injection check
- A change takes a URL from user input and SSRF is suspected
- A commit may contain real customer email addresses
Not a fit when
- Warden is not set up in the project, since the skills work only under it
- A deep manual pentest or business logic review is needed rather than a first pass by rules
- Dependency and license scanning is needed, which calls for another tool
Example request
Add the wrdn-authz skill to warden.toml and review this branch for missing ownership checks and privilege escalation.Limitations
The skills only describe what to look for, and Warden does the work: it provides the run, cache and findings output. Remote skills are attached through `warden.toml` with `warden add --remote getsentry/warden-skills --skill <name>`, and a version can be pinned with `@<ref>`. Findings should be checked by hand: the rules catch known classes of bugs and give false positives. The `vercel-deepsec` skill is meant for benchmarking coverage and open-ended scanning, not for fine tuning of a project. The full workflow is described in the Warden docs.
How to disable. Remove the skill from `warden.toml` and refresh the cache with `warden sync`. The repository itself leaves nothing on the machine.
FAQ
What is Warden?
A Sentry tool that runs skills written in Markdown on every change and shows findings as suggested edits in a pull request.
How do I attach a skill?
With `warden add --remote getsentry/warden-skills --skill <name>`, which adds a remote skill to your project `warden.toml`.
How do I pin a version?
Add a ref after the repository name: `getsentry/warden-skills@<ref>`. `warden sync getsentry/warden-skills` refreshes the cache.
Which classes of bugs are covered?
Authorization, code execution, data leaks, denial of service, GitHub Actions workflows and personal data, plus the broad `vercel-deepsec` scan.
Can skills be tested locally?
Yes, the test loop is in the repository `TESTING.md` and contribution rules are in `CONTRIBUTING.md`.
Discussion
No questions or reviews yet.
Related
SkillsA security scanner for AI agent skills: static analysis, LLM review and runs of suspicious skills in a Docker sandbox.
MCP server and CLI that check commands, paths, replies and MCP tool definitions, plus a scan for Chinese AI compliance risks
A secure gateway between AI agents and enterprise systems: identity, policy and audit instead of long-lived keys in the agent.
A security scanner for agent skills: finds prompt injection, data exfiltration and malicious code before you install
A skill for reviewing code for vulnerabilities against OWASP Top 10:2025, ASVS 5.0 and the 2026 lists for LLM and AI agent security
Cybersecurity Skills
Cybersecurity Skills for AI Agents
42 security skills for Claude Code, Cursor and Codex: code and cloud audits, authorization-gated pentesting and incident response
README in short
Each skill folder contains one `SKILL.md` with a scope. Warden executes, and the skill defines what counts as a problem.