All 92 articles, sorted alphabetically
Practical LLM Security
Implementation guides, worked examples, and interactive labs for protecting AI applications from abuse and fraud: the 25-chapter abuse and fraud prevention series, plus RAG and Knowledge Security and Multi-Agent Security.
Explore Practical LLM Security →Netty Security Updates Close an mTLS Session Resumption Bypass
Netty 4.2.20.Final and 4.1.140.Final, released on 2 November 2026, fix ten and eight advisories. How a TLS session minted without a client certificate could be resumed on a context that requires one, what the DNS and HTTP fixes change, what a local check of the jars showed, and what to upgrade.
Read article →Path Traversal and Zip Slip in File-Handling Code: Resolve, Then Check Containment, and the Cases That Break Naive Filters
How file paths built from user input or archive entries escape their directory, why string prefix checks and dot-dot filters fail, a resolve-then-contain check in Python and Java, what Python's zipfile and tarfile filters do, and tests for the cases that matter.
Read article →Vault 2.2 Security Fixes: Case-Variant Names That Slipped Past Deny Rules, Trailing-Slash LIST Bypass and What to Audit
HashiCorp Vault 2.2.0, released 15 October 2026, fixes a family of authorization bugs where case-variant names and a trailing slash could evade deny rules. This reads the changelog, explains why a deny on super-admin did not stop Super-Admin, and lists what to audit after upgrading.
Read article →Verifying JWTs Through a Signing-Key Rotation: JWKS Caching, Unknown kid Refetch, Overlap Windows and Algorithm Pinning
When an issuer rotates its signing key, tokens signed with the new key arrive before your cached keys know about it, and verification fails for valid tokens. This explains the kid header, JWKS caching with a bounded refetch, the overlap window, and the algorithm-confusion trap to avoid.
Read article →Node.js Security Releases of 21 October 2026: Non-Canonical IP Hosts in TLS Checks, NUL Bytes in DNS Names, and Repeated Random Values in Snapshots
What the Node.js 26.12.1, 24.22.1 and 22.23.4 security releases change: TLS hostname checks that now reject non-canonical IP spellings, DNS resolver calls that reject NUL bytes, random caches cleared in startup snapshots, a keep-alive socket guard, and how to triage and test a fleet.
Read article →curl 8.23.0 Fixes 24 Vulnerabilities: One High-Severity HTTP/2 Use-After-Free, a Cluster of Credential-Scoping Bugs, and How to Triage libcurl in Your Fleet
curl 8.23.0, released 14 October 2026, fixes 24 vulnerabilities. One is rated high: an HTTP/2 upload use-after-free. Many of the low ones send credentials or reuse authenticated connections across origins. Which apply to you depends on features, so triage by usage, not by headline count.
Read article →Istio Security Release 2026-007: an AuthorizationPolicy Method Bypass Over HTTP/2, an istiod Crash in Ambient Mode, and the End of 1.29
On 15 October 2026 Istio shipped 1.31.2, 1.30.6 and 1.29.9 for ISTIO-SECURITY-2026-007: a case-sensitive method matcher that lets HTTP/2 clients bypass DENY rules, a ServiceEntry that crash-loops istiod, and two Envoy crashes. What to check, how to mitigate, and why 1.29 clusters must move now.
Read article →PgBouncer 1.26.1 Fixes a SCRAM Authentication Bypass: Triage, Patch and Verify
PgBouncer 1.26.1, released 14 October 2026, fixes an authentication bypass and two denial-of-service bugs. Affected versions and preconditions from the changelog, a triage table, rolling-restart patching now that online restart is gone, and verification steps.
Read article →Refresh Token Rotation and Reuse Detection: Implementation, Races and Failure Analysis
How to implement OAuth refresh token rotation with reuse detection as RFC 9700 describes: token families, an atomic rotate operation, the concurrency and lost-response races that log users out, grace windows, and tests.
Read article →Document Ingestion Security for RAG: Connectors, Parsers, and Quarantine
Treat the ingestion pipeline as a security boundary: least-privilege connectors, isolated parsers, admission records, and a quarantine path for suspicious documents.
Read article →Permission-Aware Retrieval in Practice: Filter Pushdown, Starvation, and Row-Level Security Traps
Enforce document permissions inside the vector query, avoid filtered-search starvation, and test the database and vector-store behaviors that silently bypass access rules.
Read article →Embedding and Index Poisoning: Detect Retrieval Takeover Before It Reaches the Model
Understand how a few crafted passages can dominate retrieval, then control write paths, cap source concentration, monitor golden queries, and roll back poisoned index versions.
Read article →Citation Integrity Over Time: Versioned Spans, Revalidation, and Laundered Sources
Keep citations true after the answer is written: bind them to versioned source spans, revalidate on display, and stop summaries and agents from laundering unverifiable sources.
Read article →Deletion and Permission Revocation Across RAG Indexes and Caches
Propagate document deletion and permission changes through chunks, vector and keyword indexes, semantic caches, memory, and logs, then prove the change took effect.
Read article →Delegated Identity for Agent Chains: Token Exchange, Attenuation, and Audience per Hop
Carry the user, the acting agent, and a shrinking set of permissions through every hop of a multi-agent workflow using token exchange, audience binding, and short lifetimes.
Read article →Agent-to-Agent Trust: Verify the Peer, Pin Its Keys, and Distrust Its Output
Separate identity, capability, and output trust for remote agents: verify signed Agent Cards against pinned keys, detect card drift, and validate every artifact as untrusted data.
Read article →Confused-Deputy Failures in Multi-Agent Chains: Five Traces and the Tests That Catch Them
Diagnose how locally correct agents compose into unauthorized actions, using five concrete delegation traces and a permission-matrix test harness.
Read article →Sandboxing Agent-Generated Code in Multi-Agent Pipelines: Executors, State, and Untrusted Output
Choose and configure ADK code executors, prevent cross-user state leakage, keep secrets and networks out, and treat sandbox output as untrusted input to the next agent.
Read article →Containing a Compromised Agent: Quarantine, Credential Revocation, and Taint Tracing
Bound the blast radius of one misbehaving agent: give each agent its own identity, quarantine it without stopping the system, revoke its credentials, and trace every output it influenced.
Read article →LLM Abuse Velocity: Quotas, Campaign Signals, and Bounded Agent Work
Limit repeated misuse and runaway workloads using shared budgets, behavioral evidence, and fair enforcement.
Read article →Account Takeover and Deepfake Requests: Identity Controls for LLM Support
Keep conversational persuasion, synthetic media, and recovery claims from becoming account access or payment authority.
Read article →ADK Guardrail Callbacks: Enforce the Right Boundary Before Side Effects
Place checks around model and tool execution, return explicit denials, and test that rejected ADK calls have no side effects.
Read article →ADK Human Confirmation: Bind Approval to the Exact Action
Build confirmation flows that resist changed arguments, stale permissions, replay, and misleading approval screens.
Read article →ADK Security Plugins: Shared Policy, Ordering, and Short-circuit Hazards
Centralize cross-agent checks while testing plugin precedence, cached responses, and per-request isolation.
Read article →ADK Output Guardrails and Streaming: Prevent Disclosure Before Release
Design output validation around irreversible streaming, tool results, generated links, and authoritative action status.
Read article →ADK Tool Authorization: Bind User Identity, Tenant Scope, and Capabilities
Design narrow tools whose backend permissions remain correct even when the model proposes unauthorized arguments.
Read article →AFP Observability and Incident Response: Reconstruct Decisions Without Leaking Data
Build privacy-conscious decision traces, containment controls, and evidence-driven recovery for LLM abuse and fraud incidents.
Read article →A Production AFP Blueprint: From Read-only ADK Pilot to Controlled Actions
Combine threat modeling, guardrails, authorization, evaluation, rollout, and incident readiness into a staged deployment plan.
Read article →LLM Fraud Scoring: Base Rates, Evidence, and Human Review Capacity
Use models to organize fraud evidence without turning uncertain scores into automatic accusations or unsafe approvals.
Read article →Gemini Safety Settings and Business Guardrails: Different Scores, Different Jobs
Configure content filtering deliberately while keeping fraud, authorization, and action limits in application policy.
Read article →Guardrail Metrics: Precision, Recall, Escape Rate, and Review Burden
Choose metrics that reveal both missed harm and unnecessary friction, with explicit denominators and production assumptions.
Read article →Guardrail Outages and Fail-closed Design: Keep Useful Service Without Unsafe Actions
Define per-capability fallback behavior for classifier timeouts, policy failures, budget-store outages, and partial inspection.
Read article →Indirect Prompt Injection in RAG: When Retrieved Data Becomes an Instruction
Trace poisoned documents through retrieval, summarization, tool use, and answer rendering, then enforce source-specific authority.
Read article →How LLM Jailbreaks Happen: Instruction Confusion and Context Pressure
Understand direct jailbreak mechanisms, long-context influence, attack transfer, and what model-level defenses can and cannot prove.
Read article →LangChain Guardrails and LangGraph Approval Flows: Enforce Across Resume
Compare middleware screening with durable workflow interrupts, and prevent replay or stale approval from authorizing effects.
Read article →MCP Tool Security for AFP: Authorization, Tool Metadata, and Egress
Protect an agent from overprivileged MCP integrations, misleading tool descriptions, token misuse, and unsafe outbound requests.
Read article →Model Armor in an AFP Pipeline: Coverage, Failure Policy, and Evidence
Integrate a screening service without confusing a successful API call with complete inspection or permission to act.
Read article →Multimodal Prompt Injection: PDFs, OCR, Images, and Voice Boundaries
Trace how extracted text and synthetic media can influence an agent, and design modality-aware screening and verification.
Read article →Multi-turn Jailbreaks and Memory Poisoning: Preserve Provenance Across Time
Protect conversation state, summaries, and persistent memory from gradual privilege changes and untrusted instructions.
Read article →NeMo Guardrails and Policy-as-Code: Separate Dialogue Rules from Authority
Design input, retrieval, dialogue, execution, and output controls with explicit policy ownership and testable business invariants.
Read article →Preventing Outbound Abuse: Consent, Recipient Scope, and Agent Messaging
Keep email, messaging, and document-sharing tools from turning an LLM assistant into an unapproved bulk-action service.
Read article →Red-team Evaluation for AFP: Test Attempts, Effects, and Legitimate Use
Build authorized synthetic evaluations that distinguish model violations, blocked tool calls, and actual business harm.
Read article →Refund Fraud in LLM Agents: Atomic Reservations, Approval Binding, and Replay
Protect refund workflows from persuasive claims, concurrent requests, changed destinations, and uncertain processor outcomes.
Read article →Why LLM Abuse and Fraud Prevention Needs a Threat Model
Map attacker goals, trust boundaries, business losses, and enforceable controls before selecting an LLM guardrail.
Read article →JWT vs Session Tokens, in depth: reference versus self-contained tokens, revocation, browser storage and the hybrid most systems need
A decision guide to JWTs and opaque session tokens: what each actually is, measured token sizes, per-request cost, the revocation problem and four way…
Read article →Kubernetes Pod Security Standards, in depth: Baseline and Restricted control by control, Pod Security Admission modes, versions, exemptions and rollout
How the Kubernetes Pod Security Standards work: every Baseline and Restricted control, Pod Security Admission enforce, audit and warn modes, version p…
Read article →LLM Abuse Prevention: Design a Gateway, Budgets, and Incident Response, in depth
Design an LLM abuse gateway: ordered checks with timeout behaviour, cost-unit budgets, atomic reservations in Redis, job budgets across retries, a gra…
Read article →Fraud Prevention in LLM Applications: Verify Identity, Evidence, and Actions, in depth
Stop LLM support workflows from paying fraudsters: claim states and provenance, locked refund reservations, MAC-bound approvals, out-of-band destinati…
Read article →LLM Jailbreaks and Prompt Injection: Build Defenses at the Trust Boundary, in depth
Defend LLM applications against jailbreaks and prompt injection at the trust boundary: origin labels and taint, capability gating, plan-then-execute, …
Read article →mTLS for Internal Services, in depth: workload identity, authorizing the peer, permissive-to-strict rollout, library versus sidecar and handshake debugging
How to run mutual TLS between internal services: what it does and does not protect, SPIFFE workload identities in URI SANs, authorizing the peer ident…
Read article →mTLS for Service-to-Service, in depth: the client-auth handshake, verification versus authorization, and identity across proxies and pools
How mutual TLS works between services on the wire and across infrastructure: the TLS 1.3 CertificateRequest flight, what verifiers check, Go server an…
Read article →OAuth2 PKCE Flow Explained, in depth: verifier, challenge, the token exchange and the downgrade defence
A step-by-step guide to OAuth 2.0 PKCE (RFC 7636) and current best practice (RFC 9700): generating the verifier and S256 challenge with a test vector,…
Read article →OWASP API Security Top 10, in depth: BOLA, property and function level authorization, resource limits, SSRF and API inventory
A practical guide to the OWASP API Security Top 10 2023: why it differs from the web list, where each fix belongs in the request path, code for object…
Read article →SBOMs and Supply Chain Security, in depth: SPDX versus CycloneDX, purl identifiers, where to generate, VEX, signed attestations and answering are-we-affected in minutes
A practical guide to Software Bills of Materials: SPDX and CycloneDX, purl versus CPE matching, build versus image SBOMs and their blind spots, VEX, s…
Read article →Authentication, in depth: password storage, login defences, MFA and recovery
How to build authentication that resists real attacks: factors and phishing resistance, Argon2id password storage with OWASP parameters, NIST SP 800-6…
Read article →CSRF, in depth: how cross-site request forgery works, the variants that bypass defences, SameSite and Fetch Metadata
An attacker's-eye guide to cross-site request forgery: ambient authority, site versus origin, why CORS does not stop simple requests, JSON CSRF v…
Read article →JWT, in depth: token anatomy, signing algorithms, claims design and key rotation
How JSON Web Tokens work from the issuer's side: a token decoded by hand, JWS versus JWE, choosing HS256, RS256, ES256 or EdDSA, claims design wi…
Read article →OAuth 2.0, in depth: roles, grants, tokens and the security rules that make it safe
OAuth 2.0 from first principles: the four roles and their endpoints, the authorization code exchange on the wire, choosing between authorization code …
Read article →OWASP Top 10 (2025): Each Risk Explained, With Fixes and Tests
The OWASP Top 10:2025 explained in depth: what changed from 2021, each category from Broken Access Control and SSRF to Mishandling of Exceptional Cond…
Read article →Secret Management, in depth: inventory, secret types in code, leak channels, CI pipelines and a revoke-first incident runbook
Secret management beyond the vault: what makes a value a secret, scope, lifetime and revocability, building a credential inventory, secret wrapper typ…
Read article →SQL Injection, in depth: dynamic SQL in the database, privilege containment and detection
SQL injection from the database side: dynamic SQL inside PL/pgSQL, T-SQL, Oracle and MySQL routines and how to bind it safely, SECURITY DEFINER and se…
Read article →Threat Modeling, in depth: running the session, a worked model of a webhook service, SSRF and signing, ranking, threats as code and keeping the model current
A practical guide to running threat modeling: the four questions as a session agenda, a full worked model of a webhook delivery service, SSRF and HMAC…
Read article →Cross-Site Scripting (XSS), in depth: stored, reflected and DOM XSS, output contexts, sanitisation, Trusted Types and how to test for it
A practical guide to cross-site scripting: why it happens, the stored, reflected and DOM-based classes, encoding by output context, a worked vulnerabl…
Read article →Secret Rotation in Kubernetes, in depth: how a new credential reaches a running pod, the overlap window, and the tools that automate it
How secret rotation really works in Kubernetes: the propagation chain from secret store to Kubernetes Secret to kubelet to pod file to process to open…
Read article →Secrets Management for Developers, in depth: identity instead of secret zero, a provider interface, caching with refresh, dynamic database credentials and local development
The developer's side of secrets management: replacing a bootstrap token with workload identity, a provider interface over Vault, files and enviro…
Read article →TLS certificate pinning architecture
Deep-dive on TLS certificate and public-key pinning: the client rejects any certificate whose key is not on a pre-shipped pinset, even one a trusted C…
Read article →Container security architecture, in depth: the threat model, hardened images, Pod Security Admission, seccomp, capabilities, user namespaces and sandboxed runtimes
A layered architecture for running containers safely: why the shared kernel is the real boundary, minimal non-root images, Pod Security Admission leve…
Read article →Content Security Policy architecture
Deep-dive on CSP: directive model and browser enforcement, why nonce/hash-based script-src with strict-dynamic beats host allowlists, frame-ancestors …
Read article →CSRF defense architecture
Deep-dive on Cross-Site Request Forgery defense: ambient authority and the identity-vs-intent gap, SameSite cookies, synchronizer and signed double-su…
Read article →Data loss prevention architecture
Deep-dive on DLP architecture: inspecting content with regex/validators, ML classifiers, and exact-data fingerprinting; classifying by data type, sens…
Read article →IAM Architecture in Depth: principals, short-lived credentials, policy evaluation and guardrail layers, traced through a real deploy
Identity and access management as a system: the components, human versus workload identity, how cloud policy evaluation actually decides a request, gu…
Read article →IAM review architecture, in depth: entitlement graphs, usage evidence, risk-ranked certification, staged revocation and drift
How to build an access review system that actually removes access: collect entitlements from every system into a graph, join usage evidence, rank by r…
Read article →JWT validation
Deep-dive on validating JSON Web Tokens correctly, built on the crucial distinction that decoding a JWT is not validating it: the header and payload a…
Read article →KMS envelope encryption architecture
Deep-dive on envelope encryption: a KMS/HSM-held customer master key that wraps per-object data keys, GenerateDataKey and the wrap/unwrap flow, local …
Read article →mTLS -- both sides prove who they are
Deep-dive on mutual TLS (mTLS): the service-to-service authentication problem, mutual authentication (both sides present and validate certs) vs regula…
Read article →OAuth 2.0 authorization code + PKCE architecture
Deep-dive on the OAuth 2.0 authorization code flow with PKCE: code_verifier/code_challenge generation and the S256 method, the front-channel authoriza…
Read article →Pentest architecture, in depth: scope and rules of engagement, methodology phases, findings data, scoring, remediation and retest
How to run a penetration test as an engineered process: how it differs from scanning, red teaming and bug bounties, scoping and rules of engagement, m…
Read article →Privileged access management architecture
Deep-dive on privileged access management (PAM): eliminating standing admin access with just-in-time, time-boxed grants; the request and approval work…
Read article →Ransomware Defense Architecture, in depth: identity choke points, segmentation, backups the attacker cannot reach, detection and the clean-room restore
A defensive architecture for ransomware: how an intrusion turns into mass encryption, hardening identity and the admin path, segmentation, backups wit…
Read article →RBAC vs ABAC architecture
Deep-dive on authorization architecture: RBAC roles and hierarchies versus ABAC attribute policies, subject/resource/action/environment context, the P…
Read article →SAML federation architecture
Deep-dive on SAML SSO: the browser as transport, assertion structure and conditions, metadata-driven certificate trust, XML signature wrapping, audien…
Read article →SAST vs DAST: Static vs Dynamic Application Security Testing
SAST vs DAST compared: static code analysis vs dynamic testing of the running app, where IAST and SCA fit, shift-left CI/CD gates, false positives and…
Read article →Secrets Scanning Architecture, in depth: detectors, entropy and verification, four checkpoints from pre-commit to history scans, triage, and remediation that revokes
How secret scanning works and how to run it: the leak lifecycle, how a detector combines keywords, regex, entropy, allowlists and live verification, w…
Read article →Session management architecture
Deep-dive on web session management: server-side sessions vs stateless tokens, cookie hardening (HttpOnly/Secure/SameSite/__Host-), per-request valida…
Read article →SIEM detection pipeline architecture
Deep-dive on SIEM architecture: delivery-guaranteed collection tiers, ECS/OCSF normalization, ingest-time enrichment, streaming rules vs windowed corr…
Read article →Software Supply Chain Security Architecture in Depth
Defensive software supply chain architecture: pinning and integrity hashes, private registries and allowlists, SBOM inventory, in-toto provenance and …
Read article →Threat modeling architecture
Deep-dive on threat modeling as a practice: data-flow diagrams and trust boundaries, STRIDE elicitation and attack trees, likelihood-impact ranking, m…
Read article →Web Application Firewall -- filtering malicious HTTP traffic
Deep-dive on Web Application Firewalls: the inspect-and-filter role, rule types (signature/anomaly/custom), OWASP coverage, positive vs negative secur…
Read article →WebAuthn and passkeys architecture, in depth: authenticators, ceremonies, credential store, policy and recovery
WebAuthn and passkeys as an architecture: platform, roaming and hybrid authenticators, synced versus device-bound keys, ceremonies, verification code,…
Read article →Zero Trust Architecture in Depth
Zero trust architecture in depth: NIST SP 800-207 policy engine, administrator and enforcement points, user, device and workload identity, Rego policy…
Read article →SQL Injection in 2026, in depth: why the parser is the problem, the variants, where parameters cannot reach, ORM escape hatches, LLM-generated SQL and how to find and eliminate the class
A first-principles guide to SQL injection as it stands in 2026: how concatenated input becomes SQL syntax, in-band, blind and second-order variants, p…
Read article →WebAuthn Passkeys Explained, in depth: the relying party view of options, authenticator data bytes, verification order, autofill sign-in and synced credentials
WebAuthn passkeys from the server side: registration and sign-in options, clientDataJSON and the authenticator data byte layout with its UP, UV, BE an…
Read article →