Topics / #adversarial-attacks

Adversarial-attacks projects on GitHub

Red-team techniques and tools for attacking AI systems, including prompt injection and jailbreaks.

As of , the most-starred Adversarial-attacks projects on GitHub are L1B3RT4S (21.9K stars), sliver (12.1K stars) and reverse-SynthID (5K stars). GitHub.Locker tracks 5 Adversarial-attacks repositories with 44K stars combined. The fastest-growing on October 17 was L1B3RT4S, with +25 stars in a day.

5Projects
44KStars in total
+61Stars in a day
Browse in the catalog →github.com/topics/adversarial-attacks ↗

Hot right now

Most stars gained in a day · Oct 17

+25

L1B3RT4S

elder-plinius/L1B3RT4S · ★ 21.9K

A public collection of jailbreak prompts — the author calls them 'liberation prompts' — meant to make flagship AI models respond without their usual refusals. It is maintained by AI red-teamer elder-plinius.

+22

Spiritual-Spell-Red-Teaming

Goochbeater/Spiritual-Spell-Red-Teaming · ★ 4K

A public collection of jailbreak prompts — the author calls them 'spells' — written to get uncensored answers out of LLMs, mainly Claude.

+9

sliver

BishopFox/sliver · ★ 12.1K

Sliver is an open source cross-platform adversary emulation and red team framework that organizations of any size can use to perform security testing.

+5

reverse-SynthID

aloshdenny/reverse-SynthID · ★ 5K

reverse-SynthID reverse-engineers Google's SynthID watermark embedded in Gemini-generated images using signal processing and spectral analysis, and ships tools that detect and surgically bypass that watermark without any access to Google's proprietary encoder/decoder.

The all-time top

5 projects · by stars

  1. #1 · elder-plinius/L1B3RT4S

    L1B3RT4S

    A public collection of jailbreak prompts — the author calls them 'liberation prompts' — meant to make flagship AI models respond without their usual refusals. It is maintained by AI red-teamer elder-plinius.

    21.9K
  2. #2 · BishopFox/sliver

    sliver

    Sliver is an open source cross-platform adversary emulation and red team framework that organizations of any size can use to perform security testing.

    12.1K
  3. #3 · aloshdenny/reverse-SynthID

    reverse-SynthID

    reverse-SynthID reverse-engineers Google's SynthID watermark embedded in Gemini-generated images using signal processing and spectral analysis, and ships tools that detect and surgically bypass that watermark without any access to Google's proprietary encoder/decoder.

    5K
  1. 4

    Spiritual-Spell-Red-Teaming

    4KGoochbeater/Spiritual-Spell-Red-Teaming

    A public collection of jailbreak prompts — the author calls them 'spells' — written to get uncensored answers out of LLMs, mainly Claude.

  2. 5

    APSD

    897hcliucs/APSD

    APSD is a large-scale dataset for studying how visually stealthy adversarial perturbations look to humans, containing 10,586 clean-adversarial image pairs with human opinion scores.

Often tagged together