What I’ve worked on. And what it taught me.
Real problems, how they were fixed safely, and the lesson each one left behind. Roles and history are on LinkedIn.
How to use this page
How to read this page
Each example is short: what broke, what was done, and what to do differently next time.
The problem
What was happening and who it affected.
The fix
What was changed, and how it was tested first.
The lesson
What to keep or prevent next time.
Skills at a glance
Five groups. Thirty seconds.
Broad range, used together — build it, secure it, keep it running, and explain it clearly.
Build
Ships working products end to end.
- React front ends
- APIs & backend functions
- Data modeling
- Third-party integrations
Secure
Finds real risk and puts fixes in order.
- Domain & host review
- Vulnerability triage
- Hardening guidance
- Incident write-ups
Operate
Keeps systems steady day to day.
- Networks & firewalls
- Identity & access
- Monitoring & alerts
- Backup & recovery
AI, responsibly
Uses AI with limits people can see.
- Practical AI features
- Prompt & review design
- Access boundaries
- Framework alignment
Work with people
Easy to work with, on any team.
- Plain-language updates
- Cross-functional delivery
- Documentation & handoff
- Stakeholder framing
The Security Prompt Library
Battle-tested prompts for security professionals and AI builders — triage logs, assess CVEs, harden systems, and secure AI agents. Browse by category, preview, and copy in one click.
Most Used
Log Anomaly Triage
Turn raw log excerpts into a prioritized list of suspicious events with reasoning.
Incident Timeline Reconstruction
Convert scattered evidence into a clean chronological incident timeline with gaps flagged.
CVE Impact Assessment
Translate a raw CVE into "does this actually affect us, and how fast must we patch?"
AI Agent Security Review
Audit an AI agent design for prompt injection, tool abuse, and data exfiltration risks.
Log Anomaly Triage
Turn raw log excerpts into a prioritized list of suspicious events with reasoning.
Phishing Email Header Analysis
Dissect full email headers to verify sender authenticity and spot spoofing.
IOC Extraction & Enrichment Plan
Pull every indicator of compromise from a threat report into a structured, actionable table.
Incident Timeline Reconstruction
Convert scattered evidence into a clean chronological incident timeline with gaps flagged.
Containment Decision Matrix
Weigh containment options (isolate, block, monitor) against business impact under pressure.
Blameless Post-Mortem Draft
Generate a structured, blameless post-mortem document from incident notes.
CVE Impact Assessment
Translate a raw CVE into "does this actually affect us, and how fast must we patch?"
Scan Result Prioritizer
Cut a noisy vulnerability scan report down to the 10 findings that actually matter.
AI Agent Security Review
Audit an AI agent design for prompt injection, tool abuse, and data exfiltration risks.
System Prompt Hardening
Stress-test and rewrite a system prompt to resist jailbreaks and instruction override.
RAG Poisoning Threat Model
Map how attackers could poison your retrieval pipeline and what to filter.
Detection Rule Builder
Turn a plain-language attack behavior into SIEM detection logic with tuning notes.
Security Automation Spec
Design a SOAR-style automation for a repetitive security task, with failure handling.
Server Hardening Checklist
Generate a prioritized, OS-specific hardening checklist scoped to a server's actual role.
Firewall Ruleset Audit
Find shadowed, overly-permissive, and stale rules in a firewall ruleset export.
Security Headers Fix Plan
Convert a security-headers scan into copy-ready header configs with rollout order.
Real problems, real fixes
Short examples from hands-on work. Filter by area or search.
0 skills found
Loading skills...
Note: These educational scenarios summarize common operational patterns. Validate commands, versions, dependencies, and rollback steps in a non-production environment before changing live systems.
Why this approach matters
Real problems, clear lessons
This library shows how issues are diagnosed, fixed, explained, and used to improve the next response.