Skip to content
Back to search

agamm

MarketplaceWorks with Claude CodeGitHub

Security skills for Claude Code

Safety scan

A static text scan, not a guarantee. Select the badge to see what was scanned and every finding.

Checked today
Badge for your README
findsafeskills safety rating

Markdown

[![findsafeskills safety rating](/api/listings/agamm-claude-code-owasp/badge.svg)](/listings/agamm-claude-code-owasp)

HTML

<a href="/listings/agamm-claude-code-owasp"><img src="/api/listings/agamm-claude-code-owasp/badge.svg" alt="findsafeskills safety rating" /></a>

Install

Install instructions from the README

As written by the repo's author. findsafeskills has not verified these steps.

Install as a Claude Code plugin:

/plugin marketplace add agamm/claude-code-owasp
/plugin install owasp-security@agamm

Updating

Third-party marketplaces don't auto-update by default. Either update by hand:

claude plugin marketplace update agamm
claude plugin update owasp-security@agamm

or turn on auto-update once: run `/plugin`, open the **Marketplaces** tab, select `agamm`, and choose **Enable auto-update**. Updates apply to the next session, or run `/reload-plugins`.

Install as a plain skill

The skill is a directory (`SKILL.md` plus on-demand `reference/` files), so install the whole folder. The easiest way is [`degit`](https://github.com/Rich-Harris/degit), which copies a GitHub subdirectory without the `.git` history:

npx degit agamm/claude-code-owasp/.claude/skills/owasp-security .claude/skills/owasp-security

Or install globally for all projects:

npx degit agamm/claude-code-owasp/.claude/skills/owasp-security ~/.claude/skills/owasp-security

What was scanned

The files the safety scan read, as written by the repository's author and shown as plain text. Lines with a finding are marked.

.claude-plugin/marketplace.json · Manifest · 16 lines

View this file on GitHub (read at the repository's current version: the scan did not record which commit it read)

  1. {
  2. "name": "agamm",
  3. "owner": {
  4. "name": "Agam More",
  5. "url": "https://github.com/agamm"
  6. },
  7. "description": "Security skills for Claude Code",
  8. "plugins": [
  9. {
  10. "name": "owasp-security",
  11. "source": "./",
  12. "description": "Security review and secure-coding skill covering OWASP Top 10:2025, ASVS 5.0, the OWASP LLM Top 10 (2026), and the OWASP Agentic Top 10 (2026).",
  13. "category": "security"
  14. }
  15. ]
  16. }
README.md · README · 165 lines

View this file on GitHub (read at the repository's current version: the scan did not record which commit it read)

  1. # OWASP Security Skill for Claude Code
  2. A Claude Code skill providing the latest OWASP security best practices (2025-2026) for developers building secure applications.
  3. ## Quick Install
  4. Install as a Claude Code plugin:
  5. ```
  6. /plugin marketplace add agamm/claude-code-owasp
  7. /plugin install owasp-security@agamm
  8. ```
  9. ### Updating
  10. Third-party marketplaces don't auto-update by default. Either update by hand:
  11. ```bash
  12. claude plugin marketplace update agamm
  13. claude plugin update owasp-security@agamm
  14. ```
  15. or turn on auto-update once: run `/plugin`, open the **Marketplaces** tab, select `agamm`, and
  16. choose **Enable auto-update**. Updates apply to the next session, or run `/reload-plugins`.
  17. ### Install as a plain skill
  18. The skill is a directory (`SKILL.md` plus on-demand `reference/` files), so install the whole
  19. folder. The easiest way is [`degit`](https://github.com/Rich-Harris/degit), which copies a
  20. GitHub subdirectory without the `.git` history:
  21. ```bash
  22. npx degit agamm/claude-code-owasp/.claude/skills/owasp-security .claude/skills/owasp-security
  23. ```
  24. Or install globally for all projects:
  25. ```bash
  26. npx degit agamm/claude-code-owasp/.claude/skills/owasp-security ~/.claude/skills/owasp-security
  27. ```
  28. ## What's Included
  29. ### Claude Code Skill
  30. Location: `.claude/skills/owasp-security/`
  31. `SKILL.md` (loaded when the skill triggers):
  32. - **Security review workflow** - a five-step checklist from entry points to report
  33. - **Finding-triage rubric** - confirm attacker-controlled input, sink reachability, and blast radius before reporting, to cut false positives
  34. - **Reporting format** - fixed finding structure (location, input-to-sink path, impact, fix, confidence) with severity rated by exploitability
  35. - **OWASP Top 10:2025** quick reference table
  36. - **OWASP Top 10 for LLM Applications (2026)** - LLM01-LLM10 risks for chatbots, RAG, and tool-calling apps
  37. - **OWASP Agentic AI Security (2026)** - ASI01-ASI10 risks for AI agent systems
  38. - **ASVS 5.0** key requirements with real 5.0 requirement IDs and levels
  39. `reference/` (loaded on demand, following Claude Code progressive-disclosure best practices):
  40. - **`review-checklist.md`** - coverage checklist for every Top 10 category (including SSRF, file handling, JWTs, CORS, and CSRF) plus LLM and agent features
  41. - **`languages.md`** - language-specific security quirks for 20+ languages with unsafe/safe examples
  42. - **`config-and-supply-chain.md`** - A02 and A03 where they actually live: Dockerfiles, Kubernetes, Terraform, framework config, security headers, lockfiles, dependency confusion, install scripts, and CI/CD workflows
  43. - **`owasp-report.md`** - deep-dive on the Top 10:2025, ASVS 5.0, the LLM Top 10 (2026), and the Agentic list (2026), with per-item attack vectors and mitigations
  44. ### Accuracy
  45. Category names, ASVS chapter structure, and ASVS requirement IDs and levels are verified
  46. directly against [owasp.org/Top10/2025](https://owasp.org/Top10/2025/),
  47. [github.com/OWASP/ASVS](https://github.com/OWASP/ASVS/tree/master/5.0/en), and
  48. the [LLM Top 10 2026 PDF](https://genai.owasp.org/resource/owasp-genai-llm-top-10-2026/) rather than paraphrased.
  49. The Agentic Top 10 names are checked against that PDF's framework-mapping appendix.
  50. This matters more than it sounds: ASVS 5.0 renumbered every chapter, so 4.0 requirement IDs
  51. do not carry over, three Top 10 categories were renamed in 2025, and the LLM Top 10 2026
  52. reordered eight of its ten entries. Much of the OWASP
  53. material circulating online still cites the old IDs and names.
  54. ## Usage
  55. Once installed, Claude Code automatically activates this skill when you:
  56. - Review code for security vulnerabilities
  57. - Implement authentication or authorization
  58. - Handle user input or external data
  59. - Work with cryptography or password storage
  60. - Design API endpoints
  61. - Build AI agent systems
  62. ### Example Prompts
  63. ```
  64. "Review this code for security issues"
  65. "Is this authentication implementation secure?"
  66. "What are the security risks in this Python code?"
  67. "Help me implement secure session management"
  68. "Check this AI agent for OWASP agentic risks"
  69. ```
  70. ## Covered Standards
  71. | Standard | Version | Focus |
  72. |----------|---------|-------|
  73. | OWASP Top 10 | 2025 | Web application vulnerabilities |
  74. | OWASP ASVS | 5.0.0 | Security verification requirements |
  75. | OWASP Top 10 for LLM Apps | 2026 | LLM/RAG/tool-calling app risks |
  76. | OWASP Agentic | 2026 | AI agent security risks |
  77. ## Language Coverage
  78. Security quirks for 20+ languages including:
  79. | Web | Systems | Mobile | Scripting |
  80. |-----|---------|--------|-----------|
  81. | JavaScript/TypeScript | C/C++ | Swift | Python |
  82. | PHP | Rust | Kotlin | Ruby |
  83. | Java | Go | Dart | Perl |
  84. | C# | | | Shell |
  85. Each language section includes common vulnerabilities, unsafe/safe code patterns, and key functions to watch for.
  86. ## Evals
  87. `evals/` holds test cases for [`claude plugin eval`](https://code.claude.com/docs/en/plugin-evals):
  88. a real finding next to a safe look-alike, safe code that should not draw High findings, a real but
  89. conditional risk that should be rated below High, an LLM agent that runs model output in a shell, and
  90. an unrelated request that should not load the skill. Each case runs with and without the skill, so the
  91. score shows what the skill adds. Evals target Sonnet and Opus; each case pins `model: sonnet`.
  92. ```bash
  93. claude plugin eval . # Sonnet, the pinned default
  94. claude plugin eval . --model opus
  95. ```
  96. Runs are real model calls billed to your account. Run the suite before and after any change to
  97. `SKILL.md` or its description.
  98. ## Alternative Installation
  99. ### Clone Full Repository
  100. ```bash
  101. git clone https://github.com/agamm/claude-code-owasp.git
  102. cp -r claude-code-owasp/.claude/skills/owasp-security YOUR_PROJECT/.claude/skills/
  103. ```
  104. ## Contributing
  105. Contributions welcome! Please:
  106. 1. Fork the repository
  107. 2. Create a feature branch
  108. 3. Run the evals, and add a case if you are fixing a behavior they missed
  109. 4. Submit a pull request
  110. The skill follows Anthropic's [skill authoring best practices](https://platform.claude.com/docs/en/agents-and-tools/agent-skills/best-practices):
  111. keep `SKILL.md` to what Claude would otherwise get wrong, and put depth in `reference/`.
  112. ## Sources
  113. - [OWASP Top 10:2025](https://owasp.org/Top10/2025/)
  114. - [OWASP ASVS 5.0](https://github.com/OWASP/ASVS/tree/master/5.0/en) — chapter files, one per V-number
  115. - [OWASP Top 10 for LLM Applications 2026](https://genai.owasp.org/resource/owasp-genai-llm-top-10-2026/)
  116. - [OWASP GenAI Security Project](https://genai.owasp.org/) — home of the LLM and Agentic lists
  117. - [OWASP Cheat Sheet Series](https://cheatsheetseries.owasp.org/)
  118. ## License
  119. MIT License - See LICENSE file for details.
  120. ---
  121. **Keywords:** OWASP, security, Claude Code, AI security, application security, ASVS, secure coding, vulnerability, injection, XSS, CSRF, authentication, authorization

Ratings and comments

Sign in to rate this skill.

Sign in

Free to use. If it’s helped, pass it on.