Commit 0d77b28
feat: add interactive show & tell demo for Conversation Guardian
- 7 scenarios: benign, escalation, offensive, feedback loop, Irregular Labs, evasion, transcript
- Interactive menu with colored terminal output (ASCII-safe for Windows)
- Demonstrates full detection pipeline with real-time scoring
- Reproduces Irregular Labs paper attack scenario -> quarantine in 3 turns
- Shows transcript audit with SHA-256 content hashing and filtered views
- Fix unicode arrow in guardian logger for Windows compatibility
Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>1 parent db450db commit 0d77b28
File tree
2 files changed
+436
-1
lines changed- docs/demos
- packages/agent-os/src/agent_os/integrations
2 files changed
+436
-1
lines changed
0 commit comments