Watchdog blocks all 609 prompt injections by checking every agent tool call against the mission
jagga_tech · reddit · 2026-10-15
A CS student built Watchdog, a deterministic (no model in the loop) guard that compares every agent tool call and result against a one-line mission before execution, failing closed. It wraps Claude Code, Codex, etc. without changes.
- promptfoo red-team (Oct 8): 414 attacks, 405 harmful tool call attempts, 0 ran; harmful actions dropped 19 → 0 over four runs
- AgentDojo: 0 of 609 injection goals succeeded across all 14 templates, no benign tasks withheld
- Cost: 2 of 29 legitimate calls refused; 22% refusal rate for in-scope calls during attack sessions
- Latency: p50 32ms, p95 64ms
Author admits no unguarded baseline or independent retest yet; raw results published at aloviaai.info. Free private beta via pip install alovia-watchdog.
More from coding & agent
- Skip the deck: present with Codex Sites and embed live demos instead — gabrielchua · 2026-10-15
- Geoffrey Huntley: hand-typing code is over—delete your skills, engineer the loop — rseroter · 2026-10-15
- AI coding agent deletes production database in 9 seconds; Keycard calls for task-based agent auth — matt_slotnick · 2026-10-15
- vLLM's bitwise-consistent RL: zero KL on Nemotron NVFP4, 91.3% throughput on DeepSeek-V4-Flash — DynamicWebPaige · 2026-10-15
- vLLM achieves bitwise-consistent RL training: DeepSeek-V4-Flash hits 0 diff over 100 steps — vllm_project · 2026-10-15
- ChatGPT Sites + MCP Events + dots: A Workflow-Triggered Agent Pattern — gabrielchua · 2026-10-15