Analyzing Code Security

Security Review Workflow

Follow these steps when conducting a manual security code review:

1. Identify the attack surface. Determine entry points: API endpoints, message handlers, file parsers, user-facing forms. Read route definitions and controller registrations to build a map.
2. Trace data flows from sources to sinks. Follow untrusted input (HTTP parameters, headers, request bodies, file uploads, external API responses) through all transformations to dangerous operations (database queries, command execution, HTML rendering, file system access).
3. Check trust boundary crossings. At every point where data crosses a trust boundary (client→server, service→service, user input→database), verify that validation, authentication, and authorization are enforced.
4. Apply framework checklists. Consult references/framework-checklists.md for OWASP Web/API/Mobile Top 10 and CWE Top 25. Check each applicable category against the code under review.
5. Adopt an adversarial mindset. Form a hypothesis (e.g., “I can bypass SSO”, “I can access another user’s vault”) and work backwards to determine what conditions would make it exploitable.
6. Map findings to CWE IDs. Every finding must include the specific CWE identifier, the code location, and the data flow that makes it exploitable.
7. Classify by practical exploitability. Distinguish between practically exploitable vulnerabilities and theoretical risks. Prioritize accordingly but document both.

Key Vulnerability Categories

The most frequently encountered categories across Bitwarden’s stack:

• Injection (CWE-89, CWE-78, CWE-77) — Unsanitized input reaching SQL queries, OS commands, or LDAP queries. Always use parameterized queries and avoid string concatenation.
• Broken Access Control (CWE-862, CWE-287, CWE-306) — Missing authorization checks, IDOR, privilege escalation. Verify per-object ownership checks and role enforcement at every layer.
• XSS (CWE-79) — User input rendered in HTML without encoding. In Angular, avoid innerHTML and bypassSecurityTrust* with untrusted content.
• SSRF (CWE-918) — User-controlled URLs in server-side requests. Validate against host allowlists.
• Insecure Deserialization (CWE-502) — Type-handling enabled on untrusted input. Avoid TypeNameHandling.All in JSON.NET.
• Path Traversal (CWE-22) — User-supplied paths reaching file system operations. Canonicalize and validate against a base directory.
• Cryptographic Failures — Weak algorithms, hardcoded keys, predictable IVs. See the reviewing-security-architecture skill for approved algorithms.

For complete framework checklists (all OWASP and CWE categories), consult references/framework-checklists.md.

For CORRECT/WRONG code examples in C#, TypeScript, and SQL, consult references/vulnerability-patterns.md.

Adversarial Review Mindset

Adopt an adversarial mindset during security code review — this differs from regular code review which seeks to strengthen code.

How to think adversarially:

1. Create a hypothesis — e.g., “I can bypass SSO”, “I can access another user’s vault”, “I can escalate from member to admin”
2. Work backwards — What conditions would need to be true for the attack to succeed? Can those conditions be fabricated?
3. Question assumptions — Is that authorization check always reached? What happens if the middleware fails? What if the token is malformed but not invalid?
4. Consider failure modes — What happens when things fail? Do they fail open (insecure) or fail closed (secure)?

Critical Rules

• Authentication before authorization. Always verify the user is who they claim to be before checking what they’re allowed to do. Never skip auth checks in “internal” endpoints.
• Validate at trust boundaries. Every point where data crosses a trust boundary (client→server, service→service, user input→database) must validate. Never trust client-side validation alone.
• Map findings to CWE IDs. Every finding must include a specific CWE identifier with evidence: the code location and the data flow that makes it exploitable.
• Practical over theoretical. Distinguish between vulnerabilities that are practically exploitable in this system vs. theoretical risks. Prioritize accordingly but document both.
• Check the whole chain. A vulnerability isn’t just the sink — trace from the source (user input) through all transformations to the sink (dangerous operation). If the chain is broken by sanitization, it’s not exploitable.

Additional Resources

Reference Files

For detailed checklists and code examples, consult:

• references/framework-checklists.md — OWASP Web Top 10, API Top 10, Mobile Top 10 (2024), CWE Top 25 lookup tables
• references/vulnerability-patterns.md — CORRECT/WRONG code examples for C#/.NET, TypeScript/Angular, and SQL