grimbard

87 skills for AI agents — anti-patterns, security patterns, static analysis tools, and code review workflows for Claude Code, Cursor, Windsurf, Gemini CLI, and 30+ other agents.

Built by a pentester and secure code reviewer. Battle-tested in real engagements.

Quick Start

npx skills add igbuend/grimbard

That's it. Your AI agent now has 79 skills covering:

• 38 security anti-patterns — detect vulnerable code (XSS, SQLi, command injection, etc.)
• 21 security patterns — implement secure designs (authentication, encryption, etc.)
• 11 static analysis tool skills — run and interpret SAST tools
• 4 review & discovery skills — SARIF triage, attack surface mapping, codebase discovery
• 3 quality assurance skills — review agents, skills, and sub-agents
• 1 ethics skill — responsible disclosure and legal frameworks

Try it

After installing, ask your AI agent:

Review this folder for security vulnerabilities

The agent automatically loads the relevant anti-pattern skills based on the code context — XSS patterns for HTML rendering, SQLi patterns for database queries, etc.

What's Included

Security Anti-Patterns (38 skills)

Each skill provides BAD (vulnerable) and GOOD (secure) code examples to help AI agents identify and fix security vulnerabilities.

Skill	CWE	Description
command-injection-anti-pattern	CWE-78	Shell command string concatenation
sql-injection-anti-pattern	CWE-89	SQL query string building
xss-anti-pattern	CWE-79	Cross-site scripting (reflected, stored, DOM)
path-traversal-anti-pattern	CWE-22	Directory traversal via user input
hardcoded-secrets-anti-pattern	CWE-798	Credentials in source code
missing-authentication-anti-pattern	CWE-306	Unprotected endpoints
missing-input-validation-anti-pattern	CWE-20	Unvalidated user input
jwt-misuse-anti-pattern	CWE-347	JWT implementation flaws
open-cors-anti-pattern	CWE-942	Overly permissive CORS
mass-assignment-anti-pattern	CWE-915	Unfiltered object binding
dom-clobbering-anti-pattern	CWE-79	HTML injection overwrites JS globals
encoding-bypass-anti-pattern	CWE-838	Validation before normalization
mutation-xss-anti-pattern	CWE-79	Sanitizer bypass via parser mutations
missing-security-headers-anti-pattern	CWE-16	Missing CSP, HSTS, X-Frame-Options
session-fixation-anti-pattern	CWE-384	Session ID not rotated after login
insecure-defaults-anti-pattern	CWE-276	Unsafe default configurations
weak-encryption-anti-pattern	CWE-327	Broken or weak crypto algorithms
weak-password-hashing-anti-pattern	CWE-916	MD5/SHA1 for passwords
unrestricted-file-upload-anti-pattern	CWE-434	Dangerous file upload handling
verbose-error-messages-anti-pattern	CWE-209	Stack traces in production
log-injection-anti-pattern	CWE-117	Unsanitized data in log entries
debug-mode-production-anti-pattern	CWE-489	Debug features in production
excessive-data-exposure-anti-pattern	CWE-200	Over-sharing in API responses
missing-rate-limiting-anti-pattern	CWE-770	No throttling on sensitive endpoints
hallucinated-packages-anti-pattern	CWE-829	AI-invented dependency names
oauth-security-anti-pattern	CWE-346	OAuth/OIDC implementation flaws
redos-anti-pattern	CWE-1333	Catastrophic regex backtracking
timing-attacks-anti-pattern	CWE-208	Non-constant-time comparisons
integer-overflow-anti-pattern	CWE-190	Arithmetic overflow/underflow
type-confusion-anti-pattern	CWE-843	Type mismatch vulnerabilities
unicode-security-anti-pattern	CWE-176	Unicode normalization attacks
ldap-injection-anti-pattern	CWE-90	LDAP query injection
xpath-injection-anti-pattern	CWE-643	XPath query injection
second-order-injection-anti-pattern	CWE-74	Stored data used unsafely later
padding-oracle-anti-pattern	CWE-649	Padding oracle cryptographic attacks
length-extension-attacks-anti-pattern	CWE-328	Hash length extension
insecure-temp-files-anti-pattern	CWE-377	Predictable temp file paths
insufficient-randomness-anti-pattern	CWE-330	Weak random number generation

Security Patterns (21 skills)

Secure design patterns from DistriNet Research. Each pattern explains when to use it, how to implement it, and what to watch out for.

Category	Patterns
Authentication	Password-based, Opaque token, Verifiable token (JWT), Session-based access control
Cryptography	Encryption, Digital signature, MAC, Key management, Crypto-as-a-service, Self-managed crypto, Cryptographic action
Data Protection	Selective encrypted storage, Selective encrypted transmission, Transparent encrypted storage, Encrypted tunnel
Access Control	Authorisation, Session-based access control
Input/Output	Data validation, Output filter
Operations	Limit request rate, Log entity actions

Static Analysis Tools (11 skills)

Skills that teach AI agents how to run and interpret results from security tools. Each skill includes installation instructions — ask your AI agent to install a tool and it will guide you through it.

Tool	Purpose
Opengrep	Pattern-based SAST (open-source Semgrep fork)
Semgrep	Pattern-based SAST
Gitleaks	Secrets and credential detection
KICS	Infrastructure-as-Code security
Noir	API endpoint and attack surface discovery
OSV-Scanner	Dependency vulnerability scanning
Depscan	Advanced SCA with SBOM/VDR
Application Inspector	Technology profiling
CodeQL	Deep cross-file static analysis
Trivy	Container and dependency scanning
SARIF Tools	Process, merge, diff, and report on SARIF files

Other Skills

Skill	Description
SARIF Issue Reporter	Triage and report SARIF findings from any tool
Attack Surface XSS	XSS-focused attack surface analysis
Codebase Discovery	Repository structure and technology mapping
Content Security Policy	CSP header analysis and bypass detection
Ethical Hacking Ethics	Legal frameworks, responsible disclosure, platform rules
Skill Reviewer	Review quality of other skills
Agent Review	Review agent configurations
Sub-Agent Review	Review sub-agent setups

Full Agent (Optional)

For the complete security review workflow with automated tool orchestration, clone the repo and use it with Claude Code:

git clone https://github.com/igbuend/grimbard.git
cd grimbard

The agent provides structured workflows:

• /grimbard-review — Full 6-phase security review (4-8 hours)
• /grimbard-quick — Quick automated scan (15-30 min)
• /grimbard-triage — Prioritize existing SARIF findings
• /grimbard-compliance — PCI-DSS, HIPAA, SOC2, GDPR audit

See agents/grimbard/AGENT.md for full documentation.

Example Review

See a complete grimbard-style security review of OWASP Juice Shop — 14 findings across P0–P3, each mapped to the grimbard skill that detects it.

DevContainer

A DevContainer is included with all security tools pre-installed — no manual setup needed. Open the repo in VS Code with the Remote Containers extension, or use GitHub Codespaces.

FAQ

Why the name grimbard?

Grimbard is the badger in the medieval fable of Reynard the Fox — a loyal supporter, defender and advisor of the cunning fox. Grimbard represents wisdom, counsel and trustworthy guidance. Perfect for a repository of security patterns and knowledge.

Grimbard also gives the advice to the wrong person. Whether that's you or the AI is for you to decide.

How do skills work?

When you ask your AI agent a security-related question, it:

1. Identifies the relevant pattern(s) based on your question
2. Loads the pattern knowledge from the SKILL.md file
3. Applies the pattern to your specific context
4. Provides implementation guidance tailored to your codebase

Should I install all skills?

Yes — they're lightweight (just markdown). The AI only loads relevant skills when needed. Having all of them available means the AI can catch more issues across your codebase.

Do these help with compliance?

Yes. The patterns cover requirements from PCI-DSS, HIPAA, GDPR, and SOC 2. However, compliance requires more than technical controls — consult compliance experts.

I found a mistake / want to improve something

Please open a GitHub issue or submit a pull request.

Roadmap

• v1.0 — All skills fully tested and validated
• v2.0 — AI agent with orchestrated tool execution
• v3.0 — The AIs will decide by then

Licensing

The code in this project is licensed under the MIT license.

The documents (e.g. markdown files) in this project are licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.

The security pattern skills were created from Security Pattern Catalogue - DistriNet Research. The Gitlab repo is licensed under a Creative Commons Attribution Non Commercial Share Alike 4.0 International License.

The anti-pattern skills were created from sec-context by Arcanum Security. The repository does not contain any copyright information (which legally means it is copyrighted by default). Awaiting clarification, but consider this work a derivative (IANAL).

Some skills are modified versions from the Trail of Bits Skills Marketplace, licensed under CC BY-SA 4.0.