Overview
This collection presents seven findings from authorized security assessments. Each finding has been validated through controlled testing and published in redacted form by CyberSecAudit.
The collection covers authorization, credential lifecycle, access control, identity, production exposure, and navigation behavior. Individual articles document the tested boundary, demonstrated impact, limitations, and remediation context.
Finding status
One High, four Medium, and two Low findings are included in the public CyberSecAudit portfolio. This count refers to the seven linked cases below—not to unrelated technical observations.
Published findings
Access Control · Administrative Privilege
Anonymous Registration Creates an Administrator
An anonymous actor could register as an administrator and perform administrative operations on researcher-owned synthetic records.
Identity · Verification Lifecycle
Verification Link Poisoning to Pre-Account Takeover
A controlled authority mutation redirected an official verification link to an audit canary before the verified account state was confirmed.
Authorization · Credential Boundary
Restricted Token Crosses an Account Credential Boundary
A restricted token could read account-level credentials and revoke one disposable credential owned by the researcher.
Lifecycle · Cascade Revocation
Descendant Credential Remains Active After Revocation
The source credential was rejected after revocation while its descendant token still produced an owner-confirmed state change.
Access Control · Project Governance
Standard Role Changes Project-Level Configuration
A standard user changed project-level configuration; the owner independently read back the durable change and restored the baseline.
Exposure · Production Artifact
Production Source Artifact Disclosure
Application source and internal workflow structure were recoverable without authentication from a production build artifact.
Navigation · URL Interpretation
Cross-Origin Redirect Through a Language Selector
A trusted-origin language selector could navigate the browser to an external origin controlled by the researcher.
Validation standard
The public portfolio describes a proof loop built around a successful baseline, one-variable mutation, matched negative control, reproduction, authoritative readback, demonstrated impact, and verified cleanup.
A scanner result, HTTP status, reflection, or model output is treated as a signal until the relevant trust boundary and resulting state are verified.
Publication boundaries
- All testing was performed within authorized scope.
- Target identities, assets, endpoints, credentials, and PII are redacted.
- Descriptions are limited to demonstrated impact.
- Raw evidence and confidential engagement details are not published.
The redacted publication is preserved on this portfolio so the research record does not depend on an external domain remaining active.