---
title: "Mayank Bansal (@monk1) | Hashnode"
description: "SWE @ Failproof AI"
url: "https://hashnode.com/@monk1"
---

@monk1

# Mayank Bansal

@monk1 · Joined July 2026

[Website](https://github.com/FailproofAI/failproofai)

## Comments

[One thing this highlights is that the risk isn't just the model it's the combination of autonomy + tools + long-running execution. A model generating bad text is one thing. An agent that can plan, use tools, and modify its environment introduces a completely different class of engineering problems. That shifts the focus toward runtime controls, least privilege, and validating actions rather than assuming prompts alone are enough](https://hashnode.com/posts/ai-misalignment-and-hacking/6a61878b67e2f58bfabbe038/comment/6a62f5cc849dd5663d6ffb21)

Comment · Article · Jul 24 · [The Summer AI Agents Started Going Rogue](https://hashnode.com/posts/ai-misalignment-and-hacking/6a61878b67e2f58bfabbe038)

[Good overview. One thing I'd add is that most production RAG failures I've seen aren't actually generation failures they're retrieval failures that look like generation failures. By the time the LLM answers incorrectly, the real bug often happened much earlier in the pipeline. That's why evaluating retrieval quality separately from answer quality has become just as important.](https://hashnode.com/posts/where-rag-fails-understand-the-limitations/6a5caa941acb29ad7971db79/comment/6a5f04b11b98234a3835a80e)

Comment · Article · Jul 21 · 2 · [Where RAG Fails: Understand the Limitations](https://hashnode.com/posts/where-rag-fails-understand-the-limitations/6a5caa941acb29ad7971db79)

[One thing I'd add is that auditability shouldn't be a reporting layer, it should be part of the execution layer. If compliance depends entirely on reconstructing events after execution, you're already relying on inference. The strongest systems I've seen make every high impact action pass through an explicit decision point where policy, intent, and runtime state are evaluated before the action is allowed to execute](https://hashnode.com/posts/compliance-is-the-boring-part-of-agentic-apps-build-it-first-anyway/6a5894929b4da6768be82816/comment/6a59d423bf39421802899629)

Comment · Article · Jul 17 · [Compliance Is the Boring Part of Agentic Apps. Build It First Anyway.](https://hashnode.com/posts/compliance-is-the-boring-part-of-agentic-apps-build-it-first-anyway/6a5894929b4da6768be82816)

[Cool. We're working on a similar space with Failproof AI, but from the runtime reliability side. It's open source, and we've found that observability and runtime guardrails complement each other really well in production. Please checkout our open source repo of Failproof AI and make the star if you like it :)](https://hashnode.com/posts/you-can-t-fix-what-you-can-t-see-the-ai-agent-observability-gap/6a5494336c2c4eafbbe66856/comment/6a5878ce825151a3b026aafd#reply-6a58847a72771640bf74caf2)

Reply · Article · Jul 16 · 1 · [You Can't Fix What You Can't See: The AI Agent Observability Gap](https://hashnode.com/posts/you-can-t-fix-what-you-can-t-see-the-ai-agent-observability-gap/6a5494336c2c4eafbbe66856)

[Really liked the distinction between "is the system up?" and "is the system right?" That's the gap we've been seeing too. One thing I'd add is that observability tells you what happened, but production systems also need a runtime layer that can stop obviously bad actions before they execute. Feels like both are becoming essential as agents move beyond demos](https://hashnode.com/posts/you-can-t-fix-what-you-can-t-see-the-ai-agent-observability-gap/6a5494336c2c4eafbbe66856/comment/6a5878ce825151a3b026aafd)

Comment · Article · Jul 16 · 1 · [You Can't Fix What You Can't See: The AI Agent Observability Gap](https://hashnode.com/posts/you-can-t-fix-what-you-can-t-see-the-ai-agent-observability-gap/6a5494336c2c4eafbbe66856)

[Ahmet Özel I like the phrase "the contract is what proves it is useful." That's exactly the mindset shift that led us to open source Failproof AI. AI can generate implementations incredibly fast, but the confidence comes from independently verifying the behavior instead of trusting the generation. Please check out our repo of Failproof AI, u will surely found it helpful](https://hashnode.com/posts/the-most-dangerous-ai-generated-code-looks-finished/6a54e1e90209698f1370e82c/comment/6a5513961ff1e4521e35f0a5#reply-6a58695304c30bcf30be5f32)

Reply · Article · Jul 16 · [The most dangerous AI-generated code looks finished](https://hashnode.com/posts/the-most-dangerous-ai-generated-code-looks-finished/6a54e1e90209698f1370e82c)

[I think we're starting to see two separate infrastructure layers emerge. One is observability capturing traces, prompts, retrieval context, and behaviour after the fact. The other is runtime enforcement policies, completion validation, and execution checks before an action happens. Neither replaces the other, but together they make production agents much easier to trust.](https://hashnode.com/posts/llm-observability-when-stack-traces-fail/6a58494b28416a9798790d34/comment/6a58596f349b1e2230c12dbf)

Comment · Article · Jul 16 · 1 · [The Behaviour Was the Bug: Observability for Systems Where Stack Traces Don't Help](https://hashnode.com/posts/llm-observability-when-stack-traces-fail/6a58494b28416a9798790d34)

[I had the same realization recently. The moment an agent gets access to real infrastructure, it stops being a prompting problem and starts being an infrastructure problem. We've been moving more and more of those decisions into runtime policies in Failproof AI instead of hoping the model remembers the rules.](https://hashnode.com/posts/the-conversation-before-the-build-aligning-with-ai-on-architecture/6a58293f43fef24b32448271/comment/6a5858a2187e6e5b5bb4f73b)

Comment · Article · Jul 16 · 2 · [The Conversation Before the Build: Aligning With AI on Architecture](https://hashnode.com/posts/the-conversation-before-the-build-aligning-with-ai-on-architecture/6a58293f43fef24b32448271)

[This is exactly the kind of thing runtime guardrails are for. I'm making Failproof AI, use policies to block access to sensitive files like .env before the agent ever gets to execute the read. It's completely open source just check out the repo of Failproof AI. Hit the star also if helpful :)](https://hashnode.com/posts/stop-committing-secrets-automatically-generate-env-example-with-gen-envex/6a57cd832a25ace6dd6c4bfa/comment/6a585802635bba76b8fbd2ed)

Comment · Article · Jul 16 · [Stop Committing Secrets: Automatically Generate .env.example with gen-envex](https://hashnode.com/posts/stop-committing-secrets-automatically-generate-env-example-with-gen-envex/6a57cd832a25ace6dd6c4bfa)

[That's been one of the biggest design goals for us. We treat false positives as bugs in the rule, not something the user should just live with. The rules are deterministic, so if a check fires incorrectly, it'll do it consistently, which makes it much easier to reproduce, debug, and either improve the rule or add a narrowly scoped exception. I'd much rather have a predictable false positive than an LLM making a different judgment every run.](https://hashnode.com/posts/the-most-dangerous-ai-generated-code-looks-finished/6a54e1e90209698f1370e82c/comment/6a576c1a54b5652ae4b5135c#reply-6a584f4bdc3eddf6a7e0932e)

Reply · Article · Jul 16 · [The most dangerous AI-generated code looks finished](https://hashnode.com/posts/the-most-dangerous-ai-generated-code-looks-finished/6a54e1e90209698f1370e82c)
