SoC Security Skills

Purpose

Routes SoC hardware security analysis requests to the appropriate specialist skill.

Classification

• If threat modeling, STRIDE, attack trees, or standards-derived threat extraction → load skills/threat-model-skill/SKILL.md
• If verification planning, SVA assertions, or security checklists → load skills/verification-scaffold-skill/SKILL.md
• If compliance mapping, gap analysis, FIPS 140-3, or ISO 21434 → load skills/compliance-pipeline-skill/SKILL.md
• If executive briefing, risk register, or board/CISO updates → load skills/executive-brief-skill/SKILL.md
• If Spectre, Meltdown, MDS, cache side-channels, or branch predictor attacks → load skills/microarch-attack-skill/SKILL.md
• If DPA, SPA, fault injection, JIL scoring, or ISO 17825 → load skills/physical-sca-skill/SKILL.md
• If kernel security, KASLR, IOMMU, seccomp, or privilege escalation → load skills/kernel-security-skill/SKILL.md
• If PQC, chiplets, UCIe, heterogeneous compute, or AI accelerator security → load skills/emerging-hw-security-skill/SKILL.md
• If TLA+, formal verification, or safety/liveness properties → load skills/tlaplus-security-skill/SKILL.md
• Default: If request involves threat identification → threat-model-skill. Otherwise ask to clarify.

Skill Registry

Load Directive

Read ONLY the relevant specialist SKILL.md based on classification above. Do not pre-load multiple specialists. If the request spans multiple specialists, complete one specialist's full procedure before loading the next.

Handoff Protocol

Pass between skills as structured data:

{
  "source_skill": "skill-name",
  "entity_type": "ThreatFindings|VerificationChecklist|ComplianceState|BriefSections|...",
  "findings": [],
  "confidence_summary": {"CONFIRMED": 0, "INFERRED": 0, "SPECULATIVE": 0}
}

Downstream skills consume upstream entity types. See shared-references/soc-security/entity-schema.md for type definitions.