Forces Unseen is an offensive cybersecurity consulting firm specializing in application, infrastructure, and novel security penetration testing.
Chicago
Joined March 2021
- Tweets205
- Following72
- Followers164
- Likes65
Pinned Tweet
Whether it's application vulnerabilities, infrastructure loopholes, or exploring the unknown in novel tech like LLMs & AI, we've got you covered.
🚀 From startups to Fortune 500s, our track record speaks volumes.
Schedule your assessment today 🔗 forcesunseen.com/contact
The risks of using environment variables for secrets
⚠️ Vulnerability Exposure
⚠️ Information Disclosure
⚠️ Limited Advantages
⚠️ Security Misconception
Thread 🧵👇
1
126
4️⃣ Security Misconception
The security community has long held a misconception about the safety of environment variables for storing secrets, overlooking their exposure to vulnerabilities.
🧵 4/5
1
2
94
5️⃣ Need for Change
Recognizing the risks, it's imperative to shift away from using environment variables for secret storage in favor of more secure and robust methods.
Find out how ephemeral filesystem mounts can be used instead! 👇
forcesunseen.com/blog/stop-s…
88
Discover vulnerabilities in LLMs, AI & more with Forces Unseen 🤖
Navigate novel tech with confidence! Schedule your assessment now 👇
forcesunseen.com/contact
86
How to filter and optimize search engines the netsec way!
🔎 The SEO problem
🔎 Introduction to Brave Goggles
🔎 The Netsec Goggle
🔎 Narwhalizer
Thread 🧵👇
6️⃣ Tailoring Searches with Narwhalizer
Narwhalizer allows for advanced customization of search parameters, like score thresholds and domain frequency, enabling users to refine their searches to match specific interests or needs.
🧵 6/6
1
54
To learn more about Brave Search, the Netsec google and Narwhalizer, check out our blog "Make Search Engines Great Again!" 👇
forcesunseen.com/blog/make-s…
44
6 techniques for manipulating LLMs for desired outcomes:
1️⃣ Persistence in Correction
2️⃣ Expanding Context for Influence
4️⃣ Leveraging Language Characteristics
5️⃣ Response Conditioning
Details in thread 🧵👇
6️⃣ Adaptive and Creative Approaches
Being adaptive and creative in prompt design and query formulation can often yield surprising and effective results in manipulating LLM outputs, especially when conventional methods fall short.
🧵 6/6
1
31
Test out these skills with doublespeak.chat! Can you discover the bot's name and make it onto the leaderboard!?
doublespeak.chat/#/
27