DEV Community

Rudratosh Shastri profile picture

Rudratosh Shastri

Backend architect, 15 years deep. Still curious, still shipping. Systems • AI Agents • Security • Breaking things on purpose 🔥

Location Indore,MP Joined Joined on  Email address rudratoshs@gmail.com Personal website https://exportbankstatement.com github website

Education

Studied Computer Science at Arya Institute of IT

Pronouns

He

Work

Technical Lead & Senior Backend Architect at PharmaForceIQ

Hacktoberfest 2026 Challenge Completion
4 Week Community Wellness Streak
Sanity Challenge Completion
Top 7
Community Favorite 4
2 Week Community Wellness Streak
Community Favorite 2
1 Week Community Wellness Streak
Writing Debut

Pin Pinned

If you got a 'DEV Support: verify your account in 12 hours' comment — stop. It's a phishing scam.
rudratosh profile

If you got a 'DEV Support: verify your account in 12 hours' comment — stop. It's a phishing scam.

#security #meta #phishing #beginners
6
Comments 5
3 min read
Can You Break Our AI Security Defenses? Help Test Buried Injections & Taintgate
rudratosh profile

Can You Break Our AI Security Defenses? Help Test Buried Injections & Taintgate

#opensource #ai #security #python
3
Comments
7 min read

Want to connect with Rudratosh Shastri?

Create an account to connect with Rudratosh Shastri. You can also sign in below to proceed if you already have an account.

Create Account
Already have an account? Sign in
How to Make RAG Show Its Evidence: Claim Verification at 15M-Document Scale

Stanford's enterprise legal RAG failure stats
rudratosh profile

How to Make RAG Show Its Evidence: Claim Verification at 15M-Document Scale

#rag #ai #llm #machinelearning
13
Comments 26
25 min read
Is Claude taking control of my development skills? I went looking for proof.

Invisible skill loss backed by Anthropic research
rudratosh profile

Is Claude taking control of my development skills? I went looking for proof.

#ai #career #productivity #programming
17
Comments 6
8 min read
Your grammar checker uploads every word you type. I built one that never leaves your Mac.
rudratosh profile

Your grammar checker uploads every word you type. I built one that never leaves your Mac.

#privacy #ai #macos #indiehackers
7
Comments 19
6 min read
Anthropic is about to be worth $2 trillion while earning 4.6x less revenue than 35 SaaS companies combined. I ran the numbers.
rudratosh profile

Anthropic is about to be worth $2 trillion while earning 4.6x less revenue than 35 SaaS companies combined. I ran the numbers.

#ai #startup #saas #business
13
Comments 6
8 min read
Naive RAG is lying to you. Here are the 7 architectures that actually survive production.
rudratosh profile

Naive RAG is lying to you. Here are the 7 architectures that actually survive production.

#rag #ai #llm #machinelearning
9
Comments
7 min read
NVIDIA will run GLM-5.3 and Kimi K3 for you, for free. I read the fine print. Here are the three catches.
rudratosh profile

NVIDIA will run GLM-5.3 and Kimi K3 for you, for free. I read the fine print. Here are the three catches.

#nvidia #ai #llm #opensource
12
Comments
10 min read
I pulled verified revenue for 5,596 startups. Here's exactly what to build in 2026 (11 benchmarks, 25 templates)
rudratosh profile

I pulled verified revenue for 5,596 startups. Here's exactly what to build in 2026 (11 benchmarks, 25 templates)

#startup #saas #indiehackers #data
11
Comments 28
16 min read
17,600 actions, no human at the wheel: how to start red teaming AI agents in 2026
rudratosh profile

17,600 actions, no human at the wheel: how to start red teaming AI agents in 2026

#ai #security #cybersecurity #beginners
8
Comments 2
13 min read
I built an AI that writes resale listings — then refuses to publish the ones it can't prove

Hacktoberfest Open-Source AI Challenge Week 2: Side Hustle Submission 💸
rudratosh profile

I built an AI that writes resale listings — then refuses to publish the ones it can't prove

#devchallenge #hf26challenge #ai #opensource
7
Comments 6
9 min read
I spent three weekends trying to secure an AI agent. Here's everything that didn't work — and the ~600 lines that did.

Why output-based guardrails fail completely
rudratosh profile

I spent three weekends trying to secure an AI agent. Here's everything that didn't work — and the ~600 lines that did.

#ai #security #opensource #llm
14
Comments 11
9 min read
I built a 200-line tool that catches an AI agent exfiltrating data — while your test says PASS
rudratosh profile

I built a 200-line tool that catches an AI agent exfiltrating data — while your test says PASS

#ai #security #opensource #llm
9
Comments 6
3 min read
Your AI agent said it blocked the attack. I read the source of 5 tools that check if it's lying.

Source code breakdown of five agent security tools
rudratosh profile

Your AI agent said it blocked the attack. I read the source of 5 tools that check if it's lying.

#ai #security #opensource #llm
19
Picked as gem Comments 62
8 min read
Can a local LLM replace Claude Code? The benchmark everyone quotes is the wrong one.

Math reveals why 95% tool reliability still fails
rudratosh profile

Can a local LLM replace Claude Code? The benchmark everyone quotes is the wrong one.

#ai #llm #coding #opensource
66
Picked as gem Comments 116
11 min read
The fake 'DEV Support' scam is back, and now it wants your card balance (a living list of every scam hitting dev.to)
rudratosh profile

The fake 'DEV Support' scam is back, and now it wants your card balance (a living list of every scam hitting dev.to)

#security #phishing #meta #beginners
9
Comments 2
7 min read
Apple won't give you an API for your Health data. A free Shortcut will (here's the whole pipeline)
rudratosh profile

Apple won't give you an API for your Health data. A free Shortcut will (here's the whole pipeline)

#ios #automation #cloudflare #tutorial
7
Comments 1
9 min read
Your GitHub profile is wallpaper. Mine is a spaceship.

Real-time data wired to SVG art
rudratosh profile

Your GitHub profile is wallpaper. Mine is a spaceship.

#github #showcase #svg #webdev
12
Comments 4
8 min read
ELI5: Why can hiding one sentence inside a web page make an AI ignore its own owner and obey a total stranger?

AI lacks an internal sense of authority
rudratosh profile

ELI5: Why can hiding one sentence inside a web page make an AI ignore its own owner and obey a total stranger?

#explainlikeimfive #ai #security #beginners
9
Comments 15
4 min read
6 Ways You're Using Claude Code Like a Chatbot — And What Tech Leads Do Instead
rudratosh profile

6 Ways You're Using Claude Code Like a Chatbot — And What Tech Leads Do Instead

#ai #productivity #programming #webdev
6
Comments 2
6 min read
Your AI guardrail is green. It's also catching nothing.
rudratosh profile

Your AI guardrail is green. It's also catching nothing.

#security #ai #machinelearning #devops
9
Comments 28
7 min read
Meta's prompt-injection detector caught 1% of real agent attacks. One config change made it 99%. That's the problem.
rudratosh profile

Meta's prompt-injection detector caught 1% of real agent attacks. One config change made it 99%. That's the problem.

#ai #security #machinelearning #opensource
7
Comments 15
6 min read
'Someone already built that' is a lie. 'It's already secure' is the dangerous one.
rudratosh profile

'Someone already built that' is a lie. 'It's already secure' is the dangerous one.

#discuss #security #webdev #career
6
Comments 2
4 min read
I let AI write my code for a month. The junior devs caught what it broke — I didn't.
rudratosh profile

I let AI write my code for a month. The junior devs caught what it broke — I didn't.

#discuss #ai #career #webdev
8
Comments 49
4 min read
Your GitHub profile shows your follower count. It also shows attackers your whole attack surface.

Pins action versions to SHAs for safety
rudratosh profile

Your GitHub profile shows your follower count. It also shows attackers your whole attack surface.

#github #security #githubactions #webdev
27
Picked as gem Comments 8
4 min read
Microsoft's own security update broke Ctrl+C in Excel. Here's the patch number, and don't uninstall the first one.
rudratosh profile

Microsoft's own security update broke Ctrl+C in Excel. Here's the patch number, and don't uninstall the first one.

#news #microsoft #productivity #office
5
Comments
2 min read
The sandbox had no internet. The AI agent got out through DNS anyway.
rudratosh profile

The sandbox had no internet. The AI agent got out through DNS anyway.

#ai #security #agents #devops
5
Comments
3 min read
Your GitHub MCP server costs 55,000 tokens before your agent reads a single word.

Prompt caching softens the brutal dollar cost
rudratosh profile

Your GitHub MCP server costs 55,000 tokens before your agent reads a single word.

#mcp #ai #llm #cost
14
Picked as gem Comments 34
3 min read
Someone trained a decision model for $104. The price isn't the interesting part — it's that it refuses to write text.
rudratosh profile

Someone trained a decision model for $104. The price isn't the interesting part — it's that it refuses to write text.

#ai #machinelearning #llm #architecture
5
Comments
3 min read
Your shopping agent reads the reviews. Attackers write the reviews. Guess who wins.
rudratosh profile

Your shopping agent reads the reviews. Attackers write the reviews. Guess who wins.

#ai #security #agents #webdev
5
Comments 3
3 min read
One prompt to GPT-6 Astra can cost $10. Here's exactly when the 1M context is worth it — and when it's a trap.
rudratosh profile

One prompt to GPT-6 Astra can cost $10. Here's exactly when the 1M context is worth it — and when it's a trap.

#ai #llm #cost #productivity
5
Comments
3 min read
Rust quietly ate my JavaScript toolchain in 2026, and my builds are 20x faster for it
rudratosh profile

Rust quietly ate my JavaScript toolchain in 2026, and my builds are 20x faster for it

#javascript #webdev #rust #typescript
5
Comments
3 min read
2025 was about typing faster. 2026 is about knowing what not to build.
rudratosh profile

2025 was about typing faster. 2026 is about knowing what not to build.

#discuss #career #ai #productivity
6
Comments 1
3 min read
AI writes the code. AI reviews the code. So who's actually on the hook when it breaks?
rudratosh profile

AI writes the code. AI reviews the code. So who's actually on the hook when it breaks?

#discuss #ai #career #productivity
5
Comments 1
3 min read
Junior developer jobs didn't disappear. The bar for them quietly moved, and nobody sent the memo.
rudratosh profile

Junior developer jobs didn't disappear. The bar for them quietly moved, and nobody sent the memo.

#discuss #career #ai #beginners
5
Comments 1
3 min read
"Pinned to a SHA" is a lie your coding agent told you. Here's the Git trick behind Plugin4Shell.
rudratosh profile

"Pinned to a SHA" is a lie your coding agent told you. Here's the Git trick behind Plugin4Shell.

#security #ai #git #devops
6
Comments 19
3 min read
An AI Correctly Ignored a Forum Rumor. I Removed One Label and It Paid Out $150.

Kaggle Benchmarking Challenge Submission
rudratosh profile

An AI Correctly Ignored a Forum Rumor. I Removed One Label and It Paid Out $150.

#devchallenge #kagglechallenge #ai #machinelearning
5
Comments 4
7 min read
I Almost Built a Startup This Week. The Evidence Said No — In One Afternoon.
rudratosh profile

I Almost Built a Startup This Week. The Evidence Said No — In One Afternoon.

#discuss #startup #career #ai
6
Comments 20
5 min read
10 Years In, Everything I Was Proud Of As a Junior Was Wrong
rudratosh profile

10 Years In, Everything I Was Proud Of As a Junior Was Wrong

#discuss #career #backend #programming
5
Comments 2
4 min read
Your AI Agent Will Follow a Stranger's Instructions. Here's How I Actually Test For It.
rudratosh profile

Your AI Agent Will Follow a Stranger's Instructions. Here's How I Actually Test For It.

#ai #security #python #opensource
5
Comments 4
5 min read
The Hidden Bill: Where AI Agent Costs Actually Come From
rudratosh profile

The Hidden Bill: Where AI Agent Costs Actually Come From

#ai #agents #llm #cost
6
Comments 13
5 min read
I tested 10 prompt-injection detectors on 629 real AI agent attacks
rudratosh profile

I tested 10 prompt-injection detectors on 629 real AI agent attacks

#ai #security #llm #opensource
5
Comments 10
4 min read
Sanity's Knowledge Base stopped 6 of 7 poisoned pages. The 7th fooled Claude Opus 5.

Sanity Challenge Path One Submission
rudratosh profile

Sanity's Knowledge Base stopped 6 of 7 poisoned pages. The 7th fooled Claude Opus 5.

#devchallenge #sanitychallenge #sanity #ai
5
Comments 2
7 min read
AI agents can't tell who's giving the orders. So I built a tiny gate.
rudratosh profile

AI agents can't tell who's giving the orders. So I built a tiny gate.

#ai #security #python #opensource
12
Comments 8
7 min read
loading...