Skip to content
🆕 New skill: Security Specialist v2.0! 6-phase pipeline, 9 attack classes and adversarial validation.View skill →

Security Specialist

Full-stack application security agent — performs SAST (static analysis), DAST (dynamic testing), threat modeling, vulnerability triage, remediation and penetration testing. Combines source code review with live testing for complete evidence correlation.

  • Full security audit of a repository
  • PR/diff review for security regressions
  • Web application pentesting
  • Build threat models
  • Triage and prioritize vulnerabilities
  • Fix vulnerabilities with verified patches
  • Export findings to GitHub/Jira/Linear
  • Generate interactive HTML report
Phase What it does
1. Recon Parallel agents map architecture, trust boundaries and input surfaces
2. Hunt Parallel agents per attack class (9 categories) with 12 hunting angles
3. Validate Adversarial validation — separate agents try to DISPROVE each finding
4. Report Self-contained HTML with dark theme, filters and collapsible evidence
5. Schema findings.json validated against JSON schema (trace, conditions, execution, confidence)
6. Verify Independent verification of every factual claim by fresh agents

Injection, Access Control, Resource/File Handling, Cryptography, Business Logic, Feature Abuse, Chained Attacks, Wildcard, Obvious Things.

full-scan, diff-review, pentest, hunting, threat-model, attack-paths, discovery, triage, remediation, tracking, validation, reporting.

Each run targets gaps from prior runs. A single run finds ~50% of total vulnerabilities.

Terminal window
npx skills add https://github.com/fabricioctelles/skills -s security-specialist

v2.0 (Jun 2026)

  • 6-phase pipeline with parallel agents (inspired by Cloudflare security-audit-skill)
  • 9 attack classes and 12-angle hunting methodology
  • Adversarial validation (Phase 3) and independent verification (Phase 6)
  • JSON schema for structured findings with trace, conditions, execution, confidence
  • Zero-dependency schema validator (Node.js) for CI integration
  • Multi-run additive coverage
  • 10 anti-patterns to avoid
  • Dynamic baseline calibration in severity policy

v1.0 (Jun 2026)

  • 11 modular steering workflows
  • Python scripts (SQLite, ranker, pentest, finalizer)
  • Self-contained HTML report with dark theme
  • Pentest tool cascade
  • Three-layer correlation (source → dev → prod)

📄 Full documentation on GitHub