Palak Sethia retweeted
๐ Skills in Agent Harnesses
A lot of people don't know this (I didn't until recently) but most agent harnesses will load the description of your skill upon starting up your harness. This is how Claude Code works.
What people don't usually know is that even when you explicitly call a skill to be used inside of a harness, or the AI figures out to use a skill, it usually only does a read of the first 150 lines of the skill (to understand how to use it or if it applies to the work it's trying to do).
It's really important to outline a table in the first 150 lines of the mandatory parts of the skill that must be run when you invoke the skill.
If you make some type of mandatory instruction after 150 lines, it might get missed entirely. If you describe in prompting that you want to use a provided Python script or template, it also might be out of bounds of that 150 lines and never get called correctly.
If you write a table of contents of the skill actions in the first 150 lines and mark which ones are mandatory (or that, if the user has called the skill, you must read the whole skill context and all references), you get more reliable skills. We use markdown checklists for this now.
Palak Sethia retweeted
Learn OAuth 2.0 offsec interactively.
cristivlad25.github.io/cyberโฆ
Palak Sethia retweeted
Check out our latest tool, depx!
Seven days of malicious packages, pulled straight from OpenSSF's database and a live feed refreshed hourly.
One command replaces manual advisory tracking. You see a compromised package the day it's flagged, not after it's already in a lockfile.
Palak Sethia retweeted
Bypassing LLM security guardrails for AI red teaming usually means crafting payloads from scratch every time... ๐
P4RS3LT0NGV3 by @elder_plinius is a web-based toolkit that automatically transforms prompts using a wide range of obfuscation and encoding techniques to help test how LLMs handle adversarial inputs! ๐ค
Check it out! ๐
elder-plinius.github.io/P4RSโฆ
Palak Sethia retweeted
Dev: โCan you reproduce this?โ
Researcher: โIt works on my machine ๐
โ
Yeahโฆ thatโs not a report.
If your vuln report needs a Zoom call to explain it,
you already lost.
How to write a good vuln report (save this):
1. Describe the issue like the reader has zero context
2. Translate technical risk into business risk - who is affected and how badly?
3. Write reproduction steps so precise that anyone on the team can follow them
4. Suggest a fix - reference OWASP, a CVE, or a config change
Thatโs exactly what Barracks enforces.
Every report is structured, verified, and built to be dev-ready - so it gets fixed, not ignored.
๐ Save this before your next submission.
#Barracks #CyberSecurity #Pentesting
Palak Sethia retweeted
๐ฃ ๐ก๐๐ช ๐๐ซ๐๐ ๐๐ฅ๐ข๐ฃ โ ๐๐ด๐ฒ๐ป๐๐ถ๐ฐ ๐๐ ๐ฃ๐ฒ๐ป๐๐ฒ๐๐๐ฒ๐ฟ(๐-๐๐ด๐๐๐ฃ๐ฒ๐ป) ๐๐
๐ฎ๐บ ๐ถ๐ ๐ข๐ณ๐ณ๐ถ๐ฐ๐ถ๐ฎ๐น๐น๐ ๐๐๐ฉ๐!
๐๐ต'๐ด ๐ต๐ช๐ฎ๐ฆ ๐ต๐ฐ ๐ฃ๐ถ๐ช๐ญ๐ฅ ๐ฑ๐ณ๐ข๐ค๐ต๐ช๐ค๐ข๐ญ ๐ฆ๐น๐ฑ๐ฆ๐ณ๐ต๐ช๐ด๐ฆ ๐ช๐ฏ ๐ต๐ฆ๐ด๐ต๐ช๐ฏ๐จ ๐ข๐ฏ๐ฅ ๐ด๐ฆ๐ค๐ถ๐ณ๐ช๐ฏ๐จ ๐๐จ๐ฆ๐ฏ๐ต๐ช๐ค ๐๐ ๐ด๐บ๐ด๐ต๐ฆ๐ฎ๐ด ๐ข๐ญ๐ช๐จ๐ฏ๐ฆ๐ฅ ๐ธ๐ช๐ต๐ฉ ๐ฎ๐ฐ๐ฅ๐ฆ๐ณ๐ฏ ๐ต๐ฉ๐ณ๐ฆ๐ข๐ต ๐ญ๐ข๐ฏ๐ฅ๐ด๐ค๐ข๐ฑ๐ฆ๐ด.
๐ ๐๐ถ๐๐ฒ๐ฎ๐๐ฎ๐ ๐๐ป๐ป๐ผ๐๐ป๐ฐ๐ฒ๐บ๐ฒ๐ป๐!
Weโre offering ๐๐ฅ๐๐ ๐ฒ๐
๐ฎ๐บ ๐ฎ๐ฐ๐ฐ๐ฒ๐๐ ๐๐ผ ๐ฏ ๐๐ฒ๐น๐ฒ๐ฐ๐๐ฒ๐ฑ ๐ฝ๐ฎ๐ฟ๐๐ถ๐ฐ๐ถ๐ฝ๐ฎ๐ป๐๐
๐๐ค๐ฌ ๐ฉ๐ค ๐ฅ๐๐ง๐ฉ๐๐๐๐ฅ๐๐ฉ๐?
๐ ๐๐๐ ๐, โป๏ธ ๐๐๐๐ง๐, ๐จ๏ธ ๐พ๐ค๐ข๐ข๐๐ฃ๐ฉ & ๐ท๏ธ ๐๐๐ ๐ฎ๐ค๐ช๐ง ๐๐ง๐๐๐ฃ๐๐จ (๐๐ข๐น๐ช๐ฎ๐ถ๐ฎ ๐ฆ๐ฏ๐จ๐ข๐จ๐ฆ๐ฎ๐ฆ๐ฏ๐ต = ๐ฉ๐ช๐จ๐ฉ๐ฆ๐ณ ๐ค๐ฉ๐ข๐ฏ๐ค๐ฆ๐ด ๐ต๐ฐ ๐ธ๐ช๐ฏ!)
๐ช๐ต๐ฎ๐ ๐ถ๐ ๐-๐๐ด๐๐๐ฃ๐ฒ๐ป โ
The ๐๐ฒ๐ฟ๐๐ถ๐ณ๐ถ๐ฒ๐ฑ ๐๐ด๐ฒ๐ป๐๐ถ๐ฐ ๐๐ ๐ฃ๐ฒ๐ป๐๐ฒ๐๐๐ฒ๐ฟ (๐-๐๐ด๐๐๐ฃ๐ฒ๐ป) is an intermediate-level certification designed to test your understanding of core concepts in ๐๐ด๐ฒ๐ป๐๐ถ๐ฐ ๐๐ ๐ฎ๐ป๐ฑ ๐๐ฒ๐ป๐๐ ๐ฎ๐ฝ๐ฝ๐น๐ถ๐ฐ๐ฎ๐๐ถ๐ผ๐ป ๐๐ฒ๐ฐ๐๐ฟ๐ถ๐๐.
๐ธ Practical ๐ธ 4 Hours ๐ธ Online ๐ธ On-demand ๐ธ Real world pentesting scenarios
๐๐ค ๐ก๐ช๐๐ ๐๐ฃ ๐ฉ๐๐ ๐๐ง๐๐ฌ? ๐๐๐ง๐โ๐จ ๐ฎ๐ค๐ช๐ง ๐จ๐๐๐ค๐ฃ๐ ๐จ๐๐ค๐ฉ.ย ๐ฅ
First ๐ฎ๐ฌ๐ฌ ๐ฐ๐ฎ๐ป๐ฑ๐ถ๐ฑ๐ฎ๐๐ฒ๐ get to enroll with an exclusive ๐ด๐ฌ% ๐ฑ๐ถ๐๐ฐ๐ผ๐๐ป๐!
๐๏ธ ๐จ๐๐ฒ ๐ด๐ฌ% ๐๐ถ๐๐ฐ๐ผ๐๐ป๐ ๐๐ผ๐ฑ๐ฒ: AGENTIC-80
โ ๏ธ Once these 200 seats are filled, the offer shifts back to:
๐ณ๐ฑ% ๐ฑ๐ถ๐๐ฐ๐ผ๐๐ป๐ ๐๐๐ถ๐ป๐ด ๐ฐ๐ผ๐ฑ๐ฒ 75-OFF
๐ ๐ช๐ต๐ฎ๐โ๐ ๐ช๐ฎ๐ถ๐๐ถ๐ป๐ด ๐๐ป๐๐ถ๐ฑ๐ฒ?
โก๏ธ Multi-agent workflows โ memory, tools & reasoning chains tested
โก๏ธ LLM + API + plugin attack surface exploitation
โก๏ธ Prompt injection, tool abuse & agent hijacking covered
๐ฌ ๐ก๐ผ๐๐ฒ: The exam details will be sent to you on/before 20th April, 2026.
๐ Register here: pentestingexams.com/certificโฆ
#AgenticAI #AIPentesting #CyberSecurity #EthicalHacking #GenAI #AISecurity #RedTeaming #InfoSec #LLMSecurity #AIThreats #CyberSecurityCareer #Pentester #SecurityTesting #AIEngineering #TechCareers
Palak Sethia retweeted
I don't use net banking apps on my phone because the mandatory permissions they ask for make no sense.
Why does a banking app need access to my SMS, phone, contacts, etc., in the name of security, when not seeking invasive device permissions is, in fact, the global benchmark for cybersecurity. This is called the Principle of Least Privilege (PoLP).
โDon't do unto others what you don't want done unto youโ has been at the heart of the Zerodha philosophy.
This is exactly why we've built Zerodha the way we have. Kite asks for ZERO permissions on mobile, for instance, and this is one of the big reasons why millions of people trust us. What has enabled us is SEBI's mandatory strong two-factor authentication framework strike the right balance between security and privacy.
Palak Sethia retweeted
We've decided to have a little fun on here and run a giveaway. We are giving away a Steamdeck OLED 1TB model to one lucky individual so that they are ready in time for the release of 'The Odarian Accounts: Blood and Banner'. For you to be entered into the giveaway you must:
1. Like this post.
2. Be following the account.
3. Repost this post.
You must have done all 3 criteria to be eligible for giveaway. It will end on the 9th April 15:00 GMT. The winner will be announced on here. Weare excited to begin this and good luck. ๐๐ค โ at Gravesend, South East
Palak Sethia retweeted
Guide for learning AI security and penetration testing
github.com/anmolksachan/AI-Mโฆ
Palak Sethia retweeted
Bad News: Fake ORS Scam is BACK!
When ORSL got exposed they did not fix the product.
They just changed the name into something similar to confuse the public.
Itโs shocking that even after so much public outrage & exposing they are continuing to deceive the public.
I request all Indians to unite and support by sharing this across all platforms to make Indians aware about this deception
If we donโt call this out loudly, these brands will keep coming with new names to fool the public again.
#ORS #FoodPharmer #LabelPadhegaIndia
Palak Sethia retweeted
India is the garbage dumping capital of the world indianexpress.com/article/ciโฆ
Palak Sethia retweeted
๐ ๐๐ป๐๐ฟ๐ผ๐ฑ๐๐ฐ๐ถ๐ป๐ด ๐๐ผ๐บ๐บ๐๐ป๐ถ๐๐ ๐๐ฎ๐๐ ๐ฏ๐ ๐ง๐ต๐ฒ ๐ฆ๐ฒ๐ฐ๐ข๐ฝ๐ ๐๐ฟ๐ผ๐๐ฝย ๐
โจ ๐๐ช๐ฌ๐ฆ & ๐๐ฆ๐ด๐ฉ๐ข๐ณ๐ฆ ๐ต๐ฉ๐ช๐ด ๐ฑ๐ฐ๐ด๐ต ๐ 5 ๐๐ถ๐ค๐ฌ๐บ ๐๐ช๐ฏ๐ฏ๐ฆ๐ณ๐ด ๐ธ๐ช๐ญ๐ญ ๐จ๐ฆ๐ต ๐ข ๐๐๐๐ ๐ด๐ฆ๐ข๐ต!
At The SecOps Group, weโve always believed that cybersecurity skills and opportunities should be within everyoneโs reach. Thatโs why weโre launchingย ๐๐ผ๐บ๐บ๐๐ป๐ถ๐๐ ๐๐ฎ๐๐, ๐ผ๐๐ฟ ๐ฏ๐ถ๐ด๐ด๐ฒ๐๐ ๐๐๐ฒ๐ฝ ๐๐ฒ๐ ๐ถ๐ป ๐บ๐ฎ๐ธ๐ถ๐ป๐ด ๐ผ๐ฝ๐ฝ๐ผ๐ฟ๐๐๐ป๐ถ๐๐ถ๐ฒ๐ ๐ฎ๐ฐ๐ฐ๐ฒ๐๐๐ถ๐ฏ๐น๐ฒ ๐ณ๐ผ๐ฟ ๐ฒ๐๐ฒ๐ฟ๐๐ผ๐ป๐ฒ.ย ๐ค
๐ก ๐ช๐ต๐ฎ๐ ๐ถ๐ ๐๐ต๐ฒ ๐๐ผ๐บ๐บ๐๐ป๐ถ๐๐ ๐๐ฎ๐๐ ๐ถ๐ป๐ถ๐๐ถ๐ฎ๐๐ถ๐๐ฒ?
Community Days is a dedicated exam day where the cost is significantly reduced, seats are limited, and participants get a one-time chance to prove their skills. The exams must be taken on the event day itself; they cannot be rescheduled, and no retakes are available. Itโs designed to make exams more affordable while keeping the challenge real.
๐
๐๐
๐ฎ๐บ ๐๐ฎ๐๐ฒ: ๐ฎ๐ฒ๐๐ต ๐ฆ๐ฒ๐ฝ๐๐ฒ๐บ๐ฏ๐ฒ๐ฟ
Mark your calendars! The FIRST edition of Community Days will take place on 26th September, giving you a one-day window to take your exam at a reduced price.
๐ ๐๐
๐ฎ๐บ๐: ๐๐๐ฃ๐ฒ๐ป & ๐๐๐ฃ
Choose from two of our AppSec exams:
๐ Certified AppSec Pentester (CAPen)
Limited to 100 participants only.
Original price -ย ยฃ250
Community days price - ยฃ25 onwards
๐ Certified AppSec Practitioner (CAP)
Limited to 200 participants only.
Original price -ย ยฃ100
Community days price - ยฃ10 onwards
๐ฐ ๐ฆ๐ฎ๐น๐ฒ ๐ฆ๐๐ฎ๐ฟ๐๐: ๐ ๐ผ๐ป๐ฑ๐ฎ๐ ๐ฎ๐ฎ๐ป๐ฑ ๐ฆ๐ฒ๐ฝ๐๐ฒ๐บ๐ฏ๐ฒ๐ฟ
The Community Days sale goes live on Monday, so youโll have the chance to secure your exam seat early.
This is about creating equal opportunities for everyone to demonstrate their skills, grow, and thrive in cybersecurity.ย ๐
๐ Know more: pentestingexams.com/communitโฆ
#CommunityDays #TheSecOpsGroup #CyberSecurity #AppSec #Pentesting #CyberSkills #EthicalHacking #SkillGrowth #CyberCommunity #SecurityExams #CyberTalent #InfoSec #AffordableLearning #PentestCertification #CyberOpportunities
Palak Sethia retweeted
Prompt Injection is one of the first attack vectors used to exploit weaknesses or bypass behavior in AI models.
Here is an illustrated thread with 5 different prompt injection techniques ๐
Palak Sethia retweeted
ok, so, when in doubt for http request smuggling or http pipelining, use the bambda by @albinowax
github.com/PortSwigger/bambdโฆ
Palak Sethia retweeted
All media channels, digital platforms and individuals are advised to refrain from live coverage or real-time reporting of defence operations and movement of security forces. Disclosure of such sensitive or source-based information may jeopardize operational effectiveness and endanger lives. Past incidents like the #KargilWar, 26/11 attacks, and the #Kandahar hijacking underscore the risks of premature reporting. As per clause 6(1)(p) of the Cable Television Networks (Amendment) Rules, 2021, only periodic briefings by designated officials are permitted during anti-terror operations. All stakeholders are urged to exercise vigilance, sensitivity, and responsibility in coverage, upholding the highest standards in the service of the nation. ๐ฎ๐ณ
Read more: mib.gov.in/sites/default/filโฆ
#MediaAdvisory #NationalSecurity #MIBIndia #ResponsibleReporting
@rajnathsingh
@DefenceMinIndia
@SethSanjayMP
@HQ_IDS_India
@adgpi
@indiannavy
@IAF_MCC
@PIB_India































