To anyone using Moltbot or similar tools not confirmed security wise you can use security scanning tool for developers and startups. Vigil performs deep security audits of your system and generates cryptographically signed, tamper-evident reports. Support and appreciated!!

Hi everyone, been working on a new approach, give it a try and let me know if you can break it?? git clone https://github.com/rom-mvp/vigil.git cd vigil docker compose up --build -d

docker compose exec vigil python3 generate_api_key.py

Hey HN, maker here. We just released v2.0 of Vigil (formerly AgentShield). We kept seeing agents get blocked by security teams because they couldn't verify who the agent was or how much it was spending. v2.0 introduces a Cryptographic Identity layer. We sign requests using RSA-2048 keys (locally) or AWS KMS (in prod) so you can verify agent actions non-repudiably. We also added a hybrid cost engine that tracks token usage offline for local models (Llama 3/Mistral) so you can dev for free. Everything is Python/Flask. Would love feedback on our new 'Lazy Loading' architecture for PII redaction.

We just shipped Vigil — a plug-and-play firewall that wraps your agent in one line and blocks:

Prompt injection (DAN, GCG, context-switching)

Unauthorized tool calls

PII leaks (SSNs, cards, emails)

We’re offering a free trial for early-stage teams before we lock it behind a paywall. No setup, no sales calls just a key and go.