AI FinOps
AP2 Explained: How Google's Agent Payments Protocol Works
AP2 authorizes AI agent purchases through signed Mandates, not a single API call. Version, governance, and mechanism as of April 2026.
Blog
Insights & updates — page 6
AI agent security, identity, governance, cost, and the engineering behind the control plane.
Identity & Access
mTLS, SPIFFE, and Workload Identity for Agents
Workload identity gives an agent's runtime a cryptographic identity with no secret to steal. Where SPIFFE and mTLS fit, what they miss, and agent identity.
Identity & Access
Oasis Security Alternatives: NHI Lifecycle Management
Evaluating Oasis Security alternatives for non-human identity lifecycle management, including how it compares to Astrix and platform-native options.
AI Agent Security
Agent Sandbox Escape: Containing Code Execution
A code interpreter runs attacker-influenceable code in your infrastructure. The agent sandbox escape paths that matter and how to build one that fails safe.
AI Governance & Compliance
What FINRA and the SEC Expect From AI Agents in Finance
FINRA's 2026 oversight report and existing SEC rules set concrete expectations for AI agents at broker-dealers and RIAs — here is what they actually require.
AI Agent Security
Insecure Output Handling: Agent Output as Injection Vector
Model output is untrusted input to whatever consumes it. How agent output causes XSS, SQL injection, command execution, and log poisoning, and how to bound it.
Platform & Operations
MCP STDIO RCE Vulnerability Wave (2026) Explained
OX Security's April 2026 disclosure found a design-level flaw in Anthropic's MCP SDKs enabling remote code execution via STDIO transport.
AI Agent Security
Are Agent Skills Safe? A SKILL.md Security Checklist
Review Agent Skills before installation: inspect scripts, pin versions, test runtime permissions, and record evidence with an editable review worksheet.
Identity & Access
How to Give an AI Agent Its Own Identity
Why AI agents need first-class identity and how to model it so every action is attributable, governed, and revocable without disrupting other systems.
Platform & Operations
LiteLLM Enterprise Alternatives: Beyond Self-Hosting
When LiteLLM's self-hosted model stops scaling and what teams evaluate instead — Portkey, Kong AI Gateway, Cloudflare AI Gateway, and TrueFoundry.
Identity & Access
Sub-Agent Delegation with Scoped Tokens
An orchestrator that shares its credentials with sub-agents has no delegation model. How to pass narrowed authority down a chain, bound depth, keep attribution.
AI Agent Security
Securing AI Agents in CI/CD Pipelines
An agent in your CI/CD pipeline holds deploy credentials and reads attacker-submittable content. Why pull-request-triggered agents are the sharpest case.
AI Agent Security
WitnessAI Alternatives for Runtime AI Access Control
Evaluating alternatives to WitnessAI for runtime AI access control and compliance — what the category covers and how peers compare.
AI Governance & Compliance
US State AI Laws Compared: SB 53, RAISE, TRAIGA, Colorado
California SB 53, New York's RAISE Act, Texas TRAIGA, and Colorado's amended AI Act target different companies — here is who actually has to comply.
AI Agent Security
Voice Agent Security Risks
Voice agents add a biometric channel that can be synthesized, an audio channel that carries injection, and a real-time path with no room for review.
Platform & Operations
Agent Evaluations: Catching Quality Regressions Early
Agents regress silently when prompts, models, tools, or data change. How to build an eval harness: scoring methods, what to measure, where evals belong.
Platform & Operations
LLM Gateway Comparison 2026: Security and Governance
Portkey, LiteLLM, Kong AI Gateway, and Cloudflare AI Gateway compared for routing, spend control, and security — how to shortlist for your team.
Identity & Access
Agent Session Hijacking and Token Replay
Agent credentials live in more places than human ones and last longer. How agent sessions get hijacked, how tokens get replayed, and the binding that stops it.
AI Governance & Compliance
Governing AI Customer-Support Agents
Govern AI customer-support agents: PII detection, response guardrails, escalation triggers, and tamper-evident audit logs regulators expect.
AI Agent Security
Noma Security Alternatives for Agentic AI Security
Evaluating alternatives to Noma Security for agentic AI security — red-teaming, agent containment, and where platform-bundled options fit.
AI FinOps
Anomaly Detection on Agent Spend
Static budgets catch the disaster and miss the drift. Which agent spend signals are worth alerting on, why agents are easy to baseline, and how to avoid noise.
Platform & Operations
Agent Topology Maps: Seeing the Dependency Graph
An agent estate is a graph of agents, tools, data sources, and models. What a topology map should show, what it answers that a list cannot, how to keep it true.
AI Agent Security
Securing AI Coding Agents
AI coding agents read files, run tools, and push code autonomously. Learn the real risks — prompt injection, secret leakage — and how to contain them.
AI Agent Security
Securing Browser-Use Agents
A browser-use agent reads the open web and holds your session cookies: the highest-risk agent deployment shape in common use. Here is how to constrain it.