Stories
Concrete walkthroughs for the things people actually want — block rm -rf on Claude Code, catch a prompt injection on Codex, stop secret exfiltration from Cursor, approve risky tool calls, pin local observability, and switch connectors without losing audit history.
Six walkthroughs. Each one starts from a real operator goal, ends with a verifiable outcome, and explains the trade-offs you accepted along the way.
Stop Claude Code from rm -rf
Wire Claude Code, switch to action mode, watch a destructive shell command get blocked.
Catch a prompt injection on Codex
Local regex packs catch the obvious; the optional LLM judge catches the clever ones.
Block secret exfiltration from Cursor
Cursor's beforeShellExecution hook is the perfect stop point for `cat .env | curl ...`.
Approve risky tool calls before they fire
HITL sits between observe and full enforcement — pause, review, then continue.
Pin local observability in 60 seconds
One command brings up Prom/Loki/Tempo/Grafana wired to the gateway's OTLP endpoint.
Switch from OpenClaw to Codex without losing audit history
The audit DB is connector-agnostic. Switching only moves the data path.
Splunk
defenseclaw setup splunk runs three independent pipelines — a local Splunk in Docker for demos, remote Splunk Enterprise via HEC, and Splunk Observability Cloud via OTLP. Five demo dashboards are bundled.
Stop Claude Code from running rm -rf
Wire DefenseClaw into Claude Code, observe for a week, then promote to action mode and watch a destructive shell command never reach the disk.