Skill: java-secure-coding

Activation Contract

Use this skill when reviewing or designing Java code for secure coding practices: input validation, trust boundaries, injection, serialization/deserialization, sensitive data, resource exhaustion, mutability, access control, and third-party code risk.

Do not use this skill for full threat modeling, compliance audits, cryptography design, infrastructure hardening, or non-Java security work.

Responsibility

This skill teaches Java secure-coding review based on Java platform concerns and Oracle Secure Coding Guidelines. It does not call other skills, certify security, or replace a formal security review.

Required Context

• Data sources and trust boundaries.
• Inputs accepted and outputs generated.
• Sensitive data handled.
• Resource-intensive operations.
• Serialization, reflection, XML, SQL, command execution, or third-party dependencies involved.

Context Budget

• Keep this SKILL.md focused on secure-coding decisions.
• Use references/java-secure-coding-guidance.md for checklist details.

Hard Rules

• Establish trust boundaries before judging input safety.
• Validate and canonicalize untrusted input before use.
• Avoid dynamic SQL; use parameterized statements or safe higher-level APIs.
• Do not log secrets or expose sensitive data in exceptions.
• Release resources and defend against resource exhaustion.
• Treat deserialization, XML, reflection, JNDI, command execution, and third-party code as high-risk areas.
• Prefer least privilege and minimal accessibility.

Decision Gates

Condition	Action
Untrusted input crosses boundary	Validate type, range, size, format, and canonical form.
SQL uses string concatenation	Replace with parameterized query shape.
XML or deserialization is involved	Require restrictive parser/filter configuration.
Sensitive data appears in logs/exceptions	Remove, redact, or sanitize.
Resource size depends on input	Add limits and overflow-safe checks.
Public/extensible surface exposes internals	Reduce accessibility or document/guard contract.

Execution Steps

1. Identify trust boundaries and untrusted data.
2. Check injection, inclusion, deserialization, and resource risks.
3. Check sensitive-data exposure.
4. Check mutability, access, and privilege boundaries.
5. Return findings with severity and remediation.

Output Contract

Return:

• Secure-coding verdict: low_risk, needs_hardening, or high_risk.
• Findings with severity and evidence.
• Recommended remediation.
• Residual risks and assumptions.
• Areas requiring formal security review, if any.

References

• references/java-secure-coding-guidance.md — Java secure-coding checklist based on official guidance themes.

Assets

• None.