damn-vulnerable-ai-agent

opena2a-org

Damn Vulnerable AI Agent is a deliberately vulnerable AI agent platform for security testing and education.

Last commit

4w ago

Branch

main

Language

JavaScript

Stars

Commits (30d)

Open issues

Open PRs

Working tree

clean

Linked todos (0)

No todos linked to this repo.

fix(telemetry): report lab usage from the server path (#70)

4w ago

Fix attack-log response attribution race under concurrent same-agent requests (#58) (#69)

1mo ago

Add agent-grid screenshot + fix Multi-Step table identifier wrapping (#68)

1mo ago

Refresh the Docker Hub overview (DOCKER_README) (#67)

1mo ago

Make the all-ports docker run the hero command (#66)

1mo ago

Clean CLI output for 0.9.2: remove em dashes, fix stale --help agent list (#65)

1mo ago

Release v0.9.2: one-port quickstart (#64)

1mo ago

Proxy Attack Lab chat through the dashboard so one port is enough (#63)

1mo ago

Simplify docker run: collapse agent ports into one 7001-7021 range (#62)

1mo ago

Remove remaining em dashes across the dashboard (#61)

1mo ago

Document dashboard storytelling with screenshots; remove em dashes (#60)

1mo ago

Expose FlightBot ports (7017/7018) in Docker config (#59)

1mo ago

Attack storytelling UX: show what leaked and how to defend (#57)

1mo ago

Fix OpenAI calls on o-series/GPT-5 models + close the smoke-test gap (fixes #55) (#56)

1mo ago