Security research

A public snapshot of my early bug bounty work: what I have reported, how submissions were classified, and the areas I keep investigating.

Security research

47

Reports

3

Confirmed

8

Informative

12

Duplicates

1

Out of scope

16

Pending

7

Not applicable

Severity snapshot

Critical 0 · High 0 · Medium 1 · Low 2

Recent work

  • 01

    Bad-debt exit ordering strands later stakers in a staking protocol

    Smart-contract logic · Pending triage · Pending

    Exit queues need ordering guarantees — whoever leaves last should not absorb everyone else's bad debt.

  • 02

    Emergency exit misattributes undistributed staking rewards

    Smart-contract logic · Pending triage · Pending

    Emergency paths must settle accounting first, or rewards silently land in the wrong place.

  • 03

    Over-permissioned client-side search keys expose non-production content

    Web applications · Low · Informative

    Client-side keys inherit every permission you forget to scope — check what they can read, not just what the UI shows.

  • 04

    Payment gateway marks underpaid orders as paid

    APIs · Medium · Confirmed

    Never trust the amount from the client — reconcile what was charged against what was owed server-side.

  • 05

    Subdomain takeover via a dangling CNAME

    Web applications · Low · Informative

    DNS outlives the services it points to — audit dangling records before someone else claims them.

  • 06

    WebSocket origin allowlist bypass (cross-site WebSocket hijacking)

    Real-time systems · Pending triage · Pending

    An origin check that can be bypassed is worse than none — it hides the real attack surface.

Research notes

All security writing

What I investigate

  • Web attack surfaces. Subdomains, exposed services, dangling DNS, client-side configuration, and production boundaries.
  • API behavior. Authorization, payment state, trust boundaries, and inputs that cross application layers.
  • Real-time systems. WebSocket origin checks, session handling, and the assumptions around browser connections.
  • Smart-contract logic. Exit ordering, reward accounting, and state transitions where one user action affects another.

How I work

  1. 01

    Map the surface

    Understand scope, assets, entry points, and trust boundaries before testing.

  2. 02

    Validate the behavior

    Reproduce the smallest meaningful security impact without unnecessary access or traffic.

  3. 03

    Write useful evidence

    Document the request shape, result, impact, and a short path for triage.

Related

I keep public notes focused on the lesson rather than private program details. Read the related security research writing.