Anik 🛠️

Human | Bug Bounty Hunter 🐛| Ethical Hacker |Cyber security enthusiast | IOT and SBC Lover

[Anik 🛠️ @xettabyte1]

Hell yeah, my toys are different than yours 😅

[Roan @RohOnChain]

Met a guy making $1.4 million a year as a prompt engineer.

I asked him how he learned prompting so well.

He sent me a video that was never supposed to get out. Andrew Ng's 2 hour prompting course.

You wont find anything better about prompting than this video.

I watched it last night.

Halfway through, I realized I have been using Claude completely wrong for years.

Bookmark and watch this today.

[Isra @israfill]

you can build production AI agents with GPT-5.5, grok 4.20, AND kimi k2.6 - 500 runs/month for FREE 😳

no credit card, google login works.

stackai were acquired by asana last year and just opened up their free tier

what you get for $0:
- 500 agent runs per month (resets monthly)
- GPT-5.5, grok 4.20, kimi k2.6, claude, gemini, 30+ model providers
- visual drag-and-drop workflow builder (no code needed)
- RAG from documents, web, google drive, notion
- multi-modal: vision, text-to-speech, speech-to-text
- logic nodes: python, javascript, code execution
- browser extension, slack bot, REST API access
- 2 projects, 1 seat

what sets it apart from other free agent builders:
- founded by MIT PhDs, backed by $16M series A
- acquired by asana - not a random startup
- 100+ enterprise integrations (salesforce, sharepoint, snowflake)
- human-in-the-loop oversight
- SOC 2 / HIPAA / GDPR compliance even on free tier
- switch models per step in your workflow

how to get started (3 min):

> go to https://t.co/5fuoYSgXWe
> sign up with google - no credit card
> create a new project
> pick your model (GPT-5.5, grok, kimi, claude, whichever)
> build your agent with the drag-and-drop editor
> publish and use the chat UI or API endpoint

important:
- 500 runs/month limit - fine for testing and prototyping
- 2 projects cap - enough to experiment
- you can use temporary emails for multiple accounts to extend runs
- no production SLA, this is for building and learning
- runs reset monthly, not daily

visual agent builder + 5 frontier models + enterprise integrations = $0

while everyone else pays $20/mo for each model subscription

bookmark this before the free tier changes

[Rohit Ghumare @ghumare64]

If you want to go from AI beginner → AI builder & engineer, don’t just watch tutorials.

Build from great open-source repos.

Here are 17 free GitHub repos to learn AI/ML, LLMs, agents, RAG, math, and RL 👇🧵

[Rohit Ghumare @ghumare64]

Claude fable 5 was able to one-shot entire ai engineering from scratch.
https://t.co/DV12HanJ9a

[Akshay 🚀 @akshay_pachaar]

Google just dropped a new LLM!

You can run it locally on just 8GB RAM.

Let's fine-tune this on our own data (100% locally):

[blasty @bl4sty]

is anything documented about @AnthropicAI's tokenizer for opus-4.6/4.7/4.8? I guess token split/count stuff could be side-channeled out by sending a bunch of API requests and inferring it from the `usage` property in the response

[Rohit Ghumare @ghumare64]

Almost everyone is building agent harness systems the wrong way.

The default move: pick LangChain or LangGraph or the OpenAI Agents SDK, accept the loop, the tools, the memory, the orchestration, the policy engine, the credential store, the budget tracker, all of it, as one decision.

Mike, wrote a long piece today on why this shape is wrong, and why every long-running agent team eventually ends up rewriting its harness from scratch.

His argument: a harness isn't one thing. It's fifteen separate concerns bundled together because the surrounding ecosystem didn't give you a way to compose them. Turn state machines, provider routing, credential vaults, policy engines, approval gates, budget trackers, hook fanout, context compaction, session trees, OpenTelemetry tracing. Frameworks ship them as one block because that was the only shape available a year ago.

It isn't anymore.

When every layer is a worker on a shared bus with a typed function contract, "build your own harness" stops meaning "fork a framework." It means swap a worker. Don't like the model catalogue? Write one that hits a live API. Don't like file-backed credentials? Plug in your secrets manager. Want approvals routed through Slack instead of a console? Add a worker that calls approval::resolve. The rest of the stack does not change.

The framework era picked a position for you and locked you in. The worker model leaves the choice in your hand.

Worth reading in full.

[ProjectDiscovery @pdiscoveryio]

Neo slashed LLM costs by 59% (and up to 70% recently) using a clever prompt caching strategy. 

Unlike standard AI chatbots that reprocess entire conversations, Neo uses a "relocation trick" to keep static system prompts and tool definitions in cache, only paying for the dynamic parts of complex, multi-step security tasks. This makes running continuous, deep-dive vulnerability assessments economically viable at a scale normal AI workflows can't match.

Check out our blog to understand how we did it:
https://t.co/tNeoxeqDDj

[Akshay 🚀 @akshay_pachaar]

AI agents don't have a security problem.

They have an identity problem.

Here's what I mean:

You've deployed 47 agent workflows across your infrastructure.

At 3 am, an agent modifies a production database. You check your logs. 47 agents had access. They all used the same service account. You can't tell which one did it.

IBM's 2025 breach report found that 97% of organizations with AI-related breaches lacked proper AI access controls.

This happens because traditional security was built around human behavior: request access, get approved, complete the task, log out.

Agents don't work that way.

They run continuously for hours or days, making independent decisions at machine speed. They don't log out. They don't pause between tasks. They just keep acting.

Microsoft projects 1.3 billion agents in enterprises by 2028. Most will inherit long-lived API tokens and overprivileged service accounts.

The fix isn't better credentials. It's giving every agent its own identity.

Teleport is building an open-source framework for exactly this. I partnered with their team to show you how it works:

↳ Each agent gets its own cryptographic identity. No shared accounts, no static keys.
↳ Access decisions happen in real-time, scoped to what the agent needs right now.
↳ Every action traces back to a specific agent. Rogue behavior is detectable.
↳ Agents connect to databases and tools through the framework, not through scattered credentials.

Deploy agents safely. See exactly what each one does.

The framework is 100% open source. I have shared the link to the GitHub repo in the next tweet.

[Anik 🛠️ @xettabyte1]

@theXSSrat what about this design add as hero section although just started

[The XSS Rat - Proud XSS N00b :-) @theXSSrat]

Bug Bounty Methodology Checklist for Web Applications (B2B Apps)
https://t.co/TzAYTjLru0

[Anik 🛠️ @xettabyte1]

@ITSecurityguard @theXSSrat Now, can you tag the students or learners you’ve taught?

[Anik 🛠️ @xettabyte1]

@theXSSrat Jealousy can’t beat you, brother. Keep doing what you’ve always done. This isn’t the first time you’ve faced this, and it won’t stop you. You still have so much to offer this community.

[Md Ismail Šojal 🕷️ @0x0SojalSec]

Real AI plans, reasons, and collaborates.
17+ agentic architectures. Full, ready-to-run code.✨

Implementation of 17+ agentic architectures designed for practical use across different stages of AI system development.

- https://t.co/iXdNItUsuc

#infosec #aiagent #bugbountytips

[Matt Dancho (Business Science) @mdancho84]

This 277-page PDF unlocks the secrets of Large Language Models.

Here's what's inside: 🧵

[Akshay 🚀 @akshay_pachaar]

I boosted my AI Agent's performance by 184%

Using a fully open-source technique.

Here's the full breakdown (with code):

[Anik 🛠️ @xettabyte1]

@immunefi Hi

[Anik 🛠️ @xettabyte1]

@NamanyayG done

[Namanyay @NamanyayG]

I'm giving away Claude Code for FREE.

Yes, you read that right -- the world's best AI, now completely FREE.

My mission is to get the next billion people to write great software. I believe EVERYONE should have access to the best AI, but so many are constrained by costs.

I'm solving this by:

disrupting the world's fastest growing industry = AI

+ with the most proven business model = Ads

You can actually try it out now -- just one line of code, and you unlock all the powers of Claude Code, for free.

how it works:

1. sign up on Giga AI Free

2. change Claude Code env variables to use our inference endpoint

2. to support the free usage, we'll add ads inside the AI responses (only contextual and relevant ads, no spam)

Since we get too many signups, there's a waitlist, but we'll be rolling it out to the maximum people as we can.

Watch the video to get started & comment and retweet to get access first.

[Akshay 🚀 @akshay_pachaar]

Finally! A Text-to-SQL tool that actually works!

Vanna is an open-source RAG framework for complex Text-to-SQL generation. It manages dynamic data and allows custom RAG model training for greater accuracy.

100% open-source.