AGENT SKILL

secure-development

Use when designing, writing, reviewing, or testing code and the task calls for secure coding, security testing, threat modeling, or security review guidance.

License 许可
MIT
最近更新
2026-10-08
Stars
★ 0
Forks
0

01 前言 Front Matter

name
secure-development
license
MIT
metadata
{ "owner": "Harley Bartles", "status": "active", "use_when": [ "designing, writing, or reviewing code that handles sensitive data, trust boundaries, or external input.", "selecting or interpreting security tests, threat models, or review checklists.", "a task touches OWASP Top 10, CWE, CAPEC, or NIST control categories." ], "source-id": "secure-development", "source-path": "skills/secure-development/SKILL.md", "related_skills": [ "owasp-top-ten", "risk-gates" ], "do_not_use_when": [ "the task is purely infrastructure deployment or operations.", "another more specific skill owns the task." ], "provenance-name": "Secure Development first-party skill", "source-category": "first_party" }
description
Use when designing, writing, reviewing, or testing code and the task calls for secure coding, security testing, threat modeling, or security review guidance.

02 资源文件 5

03 SKILL.md 正文

Secure Development

Use this skill to build and verify software with security built in: secure coding, testing, threat modeling, and review.

When to Use

  • Designing or reviewing code that processes external input, secrets, or crosses trust boundaries.
  • Choosing security tests, interpreting scan results, or building a test plan.
  • Modeling threats for a feature, API, or data flow.
  • Running a security review or checklist before a commit, merge, or release.

Core Pattern

  1. Secure by design: validate all input, fail safely, apply least privilege, and keep secrets out of code.
  2. Map threats: identify assets, trust boundaries, attack surface, and likely attack paths before building.
  3. Test early and often: combine static analysis, dynamic scans, dependency checks, and targeted negative tests.
  4. Review before commit: use a checklist, check against the OWASP Top 10 and relevant CWEs, and route high-risk findings through risk-gates.

Common Mistakes

  • Treating security as a final scan instead of a design constraint. → Build it in from the first sketch.
  • Trusting input that crossed a boundary. → Validate, sanitize, and parameterize at every boundary.
  • Storing secrets in source, logs, or environment variables that leak. → Use a secrets manager and scan for leaks.
  • Relying on one test type. → Layer static, dynamic, dependency, and manual review.

Load references/operational-guidance.md for deeper coverage of secure coding, testing, threat modeling, and review checklists.

04 同仓库更多 Skills