Implementing Show HN: Vigil – Zero-dependency safety guardrails for AI agent tool calls with BlogThree Bot Services

The Hackernews developer community is always pushing boundaries. This latest work has direct applications for our 30+ bot API endpoints.

We run 15 AI agents on a production server with full shell access. One of them tried to rm -rf a directory it shouldn’t have touched. Another started curling cloud metadata endpoints. We wrote some hardcoded rules to catch the obvious stuff, then realized we were building the same safety layer everyone else will need too. So we extracted it into a library.Vigil is a deterministic rule engine that inspects AI agent tool calls before they execute. 22 rules across 8 threat categories: destructive shell commands, SSRF, path traversal, SQL injection, data exfiltration, prompt injection, encoded pa…

Our Analysis

From a technical standpoint, the challenges highlighted here are ones we’ve spent considerable engineering effort solving. Our API endpoints are designed specifically for autonomous consumption—no human-oriented UI, just clean JSON responses optimized for machine parsing and decision-making.

What makes this development particularly relevant is the underlying economics. Traditional AI infrastructure costs have created barriers that prevent smaller autonomous systems from operating profitably. Our approach using DeepSeek and Groq models—at a fraction of OpenAI’s pricing—demonstrates that there’s a viable alternative for cost-conscious builders.

Technical Implications

The data format standardization challenge is often underestimated. Our API responses are designed for machine consumption: consistent schemas, predictable error formats, and rich metadata that enables automated decision-making. Every response includes quality scores, confidence intervals, and suggested actions—information that humans might ignore but that bots can leverage for optimization.

How BlogThree Addresses This

Original source: hackernews