Security issues in a Tor-based messenger
Cwtch credited my reports in its 1.17.2 security release. The fixes cover contact permissions and denial-of-service paths in several experimental features.
Noir0x63 / Security research
Application security researcher
I work in AppSec, investigating failures in authorization, business logic and applied cryptography through source review, adversarial modeling and targeted verification.
Research ↓01 / AppSec
Application-security work is grouped here: public disclosures and authorized assessments, each tied to a reviewed system, evidence trail and remediation state.
Cwtch credited my reports in its 1.17.2 security release. The fixes cover contact permissions and denial-of-service paths in several experimental features.
The project names me for reporting an IPv4-mapped IPv6 policy bypass and a claim page exposed on the health listener. Both were addressed in v0.5.3.
A source review of the public web client, shared cryptography package and Tauri shell. The project acknowledges the review and records its own remediation decisions.
The public advisory records the vulnerability and the reporter credit; the credit is stored in the advisory metadata and may not render on the public page. It describes an unauthenticated server-side request flaw in the GitHub integration callback, followed by a corrective pull request and regression tests.
Reviewed settlement validation, merchant payment destinations and custody-key handling. Findings are linked to public advisories.
Combined source review with isolated HTTP verification of RLS policies, server-side requests and the daemon's module boundary.
Reviewed participant authorization, Supabase row-level security and the encrypted-chat architecture. The selected findings include a policy that allowed an authenticated user to add themselves to another conversation, and a user-table read policy that exposed data beyond its intended boundary.
Source review and targeted verification found feed content reaching public pages without adequate escaping, allowing stored script execution, and proxy endpoints without sufficient limits on paid upstream API use.
An independent source review identified plan-assignment RPCs without intended authorization checks, a password-reset flow insufficiently bound to reset context, and desktop distribution issues. Dynamic verification was outside the completed scope; the client remains anonymized.
Defensive reconstruction of an American Express phishing campaign: delivery, kit collection, infrastructure layers, evidence limits and response actions.
Research note on an IFFOC-linked ScreenConnect access attempt, with the observed chain, infrastructure and detection guidance.
02 / Further work
Source review and verification of authorization, business logic, and applied cryptography within an agreed scope. Deliverables include evidence, impact, remediation discussion, and retesting of agreed fixes. Scope, version, access, exclusions, and timeline are documented before work begins.
03 / Contact
For authorized security assessments, research collaboration and responsible disclosure.
[email protected]