Cybersecurity engineer: practical SecDevOps, cloud hardening, AI-agent security, and automation for small teams, MSPs and security operators.
Australia · 18+ years in tech · I build small security tools that operators can run, audit and automate.
- 🤖 AI-agent security tooling: prompt linting, prompt-injection and risk checks, SARIF output, CI gates.
- ☁️ Infrastructure resilience and hardening: Linux/VPS auditing, plus GitHub and Cloudflare backup and recovery.
- 🛰️ Threat-intel and SOC enablement: OpenCTI labs, log and traffic triage, defensive workflows.
| Project | What it does | Stack | Security angle | Status |
|---|---|---|---|---|
| promptgenie | Lints, scans, tests and runs AI-agent prompts; gates them in CI | Python · GitHub Actions | Prompt-injection and agentic-risk rules, SARIF, threat model |  |
| github-backup | Mirrors every repo you own (incl. orgs) into checksummed archives; guided restore | Bash · Bats | Token never in argv, traversal-safe extraction, sha256 verification |  |
| cloudflare-backup | Exports every Cloudflare DNS zone to timestamped BIND files | Bash · curl · jq | Read-only scoped API token, cron-friendly DR | Stable |
| vps-audit | Dependency-free Linux server security and performance audit | Bash · containers | Maintained hardened fork: attested releases, container CI |  |
| poc_opencti | One-script OpenCTI threat-intel lab on Docker Compose | Bash · Docker | Generated secrets, localhost-bound services, lab-only scope | Lab / PoC |
Also in the lab: casaos-installation-deb13 (homelab installer), velocideploy (Velociraptor deployment), powershell.
Client-side security utilities. Where it makes sense, your data stays in your browser.
- PQC tester: checks whether a site negotiates post-quantum TLS key exchange.
- Phishing analyser: explainable, rule-based phishing triage for pasted emails.
- Redirect tracer: follows and explains every hop in a redirect chain.
- Network traffic triage and Windows event log triage: learn-by-doing triage of Wireshark and Security-log CSVs.
- AI agent prompt library: ten copy-ready system prompts for AI business agents.
More writing and resources: AI security · Website security baseline · All tools
- Securing AI-agent workflows: prompt policy, CI gates, least-privilege tool use
- SecDevOps automation that turns security controls into repeatable, testable pipelines
- Cloud and web hardening for small teams and MSPs (Cloudflare, AWS, WordPress)
- Practical threat-intel and SOC workflows
Security and observability: Splunk · Velociraptor · OpenCTI · Grafana · Prometheus · WAF · IAM · threat modelling · DNS security · endpoint forensics
- AI-agent security tooling and prompt-security CI
- Secure cloud and infrastructure automation
- WordPress and web hardening
- SOC / CTI enablement and lab builds
Open an issue on any repo, or get in touch via mylesagnew.com. Security reports go through each repo's private advisory link in SECURITY.md.
📊 GitHub stats
Self-generated daily by a standard-library script in this repo (scripts/generate_stats.py), so it includes private work as totals only.
Coffee addict. Surfer. Dad. Still learning. ☕🏄