
AI
How to classify MCP tools by risk (read, edit, delete)
Sort MCP tools into read, edit and delete classes with the spec's readOnlyHint and destructiveHint. Set each class yourself, since hints are not guarantees.
6 min read
Tag
62 posts · Page 2 of 6

AI
Sort MCP tools into read, edit and delete classes with the spec's readOnlyHint and destructiveHint. Set each class yourself, since hints are not guarantees.
6 min read

AI
Pin an AI agent to its approved plan by allowlisting exact tool calls, denying the rest and checking arguments in a PreToolUse hook. Claude Code steps included.
6 min read

Events
Our KubeCon + CloudNativeCon North America 2026 recap: AT&T and Kinexys by J.P. Morgan on Kyverno in production, the move to CEL and policy for AI agents.
6 min read

AI
Tool-level authorization decides which tools an agent may call. Learn how the MCP spec splits that job and how Claude Code evaluates deny, ask and allow rules.
6 min read

AI
OpenShell's docs call a maximum-access policy a boundary. The prover checks that a subagent's policy allows nothing beyond the parent's. Here is how to run it.
7 min read

AI
ABAC grants or refuses an agent's action by checking attributes of the agent, the tool, the operation and the environment. Here is how to map it to tool calls.
9 min read

AI
OpenShell agents hold placeholders, not API keys. The supervisor swaps in the real key only where policy and the profile allow it. Models are called natively.
9 min read

AI
Write an OpenShell sandbox policy in YAML: set filesystem and Landlock rules at startup, scope each network rule to one binary, then verify changes live.
9 min read

Kubernetes
OpenShell, Agent Substrate and Kubernetes Agent Sandbox solve different parts of running AI agents. Compare what each manages, enforces and requires.
7 min read

Kubernetes
Find over-privileged AI agents by testing each agent's Kubernetes service account with kubectl auth can-i --as, checking escalation paths and reading audit logs
6 min read

Kubernetes
Build AI workflows on Kubernetes that stay deterministic: collect with Jobs, gate with CEL, and give the model one bounded step with only the fields it needs.
6 min read

AI
Cross App Access (XAA) lets your identity provider decide which MCP clients reach which MCP servers, via a signed ID-JAG exchanged for an access token.
4 min read