Skip to content
Cloud Security DeskSearch
Menu

Evidence search

Search

Search titles, summaries, topics, providers, authors, and the full open-access corpus.

Results for “Evidence”

67 publications
Technical guideSource-based analysis

Investigate an Entra application through grants and sign-ins

A successful service-principal sign-in is one event in a larger sequence. Connect it to credential changes, permission grants and the resource involved.

Detection & response · Microsoft Entra · By Cloud Security Desk
Research noteSource-based analysis

Define the expiry boundary for Entra privileged access

PIM records activation and expiry, but the protected application still determines when changed authority takes effect. Review both sides of that boundary.

Identity & access · Microsoft Entra · By Cloud Security Desk
Technical guideSource-based analysis

Measure recovery by the service you can restore

Define application acceptance, recoverable data, and a complete timeline before treating a completed restore job as proof of recovery.

Resilience · AWS / PostgreSQL · By Cloud Security Desk
Technical guideSource-based analysis

Keep MCP tokens bound to the intended resource

Keep token audiences, user consent, and downstream tool authority separate when reviewing a protected HTTP MCP service.

AI systems · MCP · By Cloud Security Desk
Technical guideSource-based analysis

Test Kubernetes egress policies beyond a successful DNS lookup

Separate DNS resolution, source egress, destination ingress, and application identity when testing Kubernetes network isolation.

Workload security · Kubernetes / Cilium · By Cloud Security Desk
Technical guideSource-based analysis

Verify the CloudTrail files behind an incident timeline

Digest delivery and successful validation are different evidence states. Preserve the files, metadata and validation result needed to distinguish them.

Detection & response · AWS · By Cloud Security Desk
Technical guideSource-based analysis

Close the AWS sessions that survive identity shutdown

Removing a login does not by itself establish that issued AWS credentials have lost their authority. Build a revocation procedure that accounts for both.

Identity & access · AWS · By Cloud Security Desk
Technical guideSource-based analysis

Protect backup copies from the account that runs production

Map deletion authority, retention protection, keys, and recovery identities so a surviving backup has a usable path back to service.

Resilience · AWS · By Cloud Security Desk
Technical guideSource-based analysis

Keep private documents out of shared RAG answers

Authorize retrieved documents before they enter model context, preserve permissions on chunks, and make source access changes visible in the retrieval path.

AI systems · Azure · By Cloud Security Desk
Technical guideSource-based analysis

Roll out Kubernetes Pod Security Admission without surprises

Stage namespace enforcement around the Pods a controller will create next, with explicit policy versions, runtime checks, and narrowly owned exceptions.

Workload security · Kubernetes · By Cloud Security Desk
Technical guideSource-based analysis

Choose the S3 object events your investigation will need

CloudTrail event history is not an object-access ledger. Build selectors around the questions an investigation must answer, then test the exclusions.

Detection & response · AWS · By Cloud Security Desk
Technical guideSource-based analysis

Lock GitHub deployment trust to the job you intend

A short-lived token still needs a narrow trust decision. Review the subject, deployment environment, reusable workflow and AWS role together.

Identity & access · AWS / GitHub · By Cloud Security Desk
Research reportDesk publication

Qwen3.8-Flash-Next and GLM-5.3-Flash share a 3:1 long-context pattern

Both models replace most conventional attention layers with recurrent state and reserve sparse attention for periodic retrieval. Their differences lie in where they place capacity, how much neural computation they activate, and what their serving stacks must keep trustworthy.

AI systems · Resilience · By Umair Akbar and Ahmed Elshekh
Research reportIllustrative analysis

The permission path you didn’t review

Cross-account trust rarely fails at the obvious policy. The risk lives in the path between identities, conditions, and inherited access.

Identity & access · AWS · By Umair Akbar and Ahmed Elshekh
Visual briefIllustrative analysis

Cloud logs that never reach the SIEM

A dashboard can be healthy while the evidence behind it is incomplete. Coverage needs to be tested from event creation to searchable record.

Detection & response · AWS / Azure / Google Cloud · By Umair Akbar and Ahmed Elshekh
Technical guideIllustrative analysis

What changes when static keys disappear

Workload identity removes a secret, but it also moves trust into issuers, claims, audiences, and runtime attachment points.

Identity & access · Workload security · AWS / Azure / Google Cloud · By Umair Akbar and Ahmed Elshekh
Field noteIllustrative analysis

Five Kubernetes events your cloud trail will not explain

Cloud control-plane logs tell you who changed the cluster. They do not fully explain what happened inside it.

Detection & response · Workload security · Kubernetes · By Umair Akbar and Ahmed Elshekh
Research noteIllustrative analysis

Threat-model the system around the model

The model endpoint is one component. The consequential paths often run through retrieval stores, orchestration identities, evaluation data, and operator tools.

AI systems · AWS / Azure / Google Cloud · By Umair Akbar and Ahmed Elshekh
Technical guideIllustrative analysis

Recover the control plane before you need it

A recovery plan that depends on the compromised identity system is an aspiration. Build and test an independent administrative path.

Resilience · Identity & access · AWS / Azure / Google Cloud · By Umair Akbar and Ahmed Elshekh