Writeups
Writeups from CTF challenges and lab-based security research, focused on web exploitation and identity and cloud attack paths. By day I work on Microsoft identity and detection and response; these are offensive-side exercises.
Everything here is CTF or authorized-lab work. No real-world targets, credentials or findings are published. The originals live in GrantZero604/security-writeups on GitHub, licensed CC BY 4.0.
CTF writeups
Section titled “CTF writeups”- 401 Authorized BetaThree bugs in one HKCERT CTF 2026 challenge: a token leaked on the denial path, a published 16-bit LCG and SQLi through PyMySQL %-substitution.
- Halloween Store1st place write-up, Intigriti CTF 1026A coupon guard that remembers one code, not the set: alternate two public codes for 100% off. Intigriti CTF 1026 business-logic writeup.
- Critter GalleryA base64-wrapped parameter, a blank-response error oracle and a union-based SQL injection into a hidden table. Intigriti CTF 0926.
- Bad ReceptionStored XSS in a moderator bot: a DOMPurify ADD_TAGS gap, a script-src 'self' CSP and a same-origin JSONP gadget. Intigriti CTF 0826.
Halloween Store took first place out of 72 write-ups in Intigriti’s October 2026 challenge. It also has an interactive edition where the coupon model runs in your browser.
401 Authorized Beta (HKCERT CTF 2026) also has an interactive edition where each guard runs as a model in your browser.
Methodology
Section titled “Methodology”- OAuth2 / OIDC discovery playbookRead-only OAuth2 and OpenID Connect recon: discovery metadata, provider fingerprints, flow classification, PKCE and redirect_uri tests.