wuyoscar

Internal Safety Collapse

Name: Internal Safety Collapse
Author: wuyoscar

AI community

SOURCE https://github.com/wuyoscar/ISC-Bench

Description

We built an adversarial codespace setup. Place any AI agent into a normal workflow inside it, and the agent will fill in whatever is missing.

Installation

This entry records only its repository, not the path inside it, so there is no exact command to give. Open the source below and copy the folder into ~/.claude/skills/, or the file into ~/.claude/agents/.

README

TVD - Internal Safety Collapse in Frontier Large Language Models

[!CAUTION] Research use only. **Internal Safety Collapse (ISC)** supports red-teaming, evaluation, and mitigation research. Do not use these materials to cause harm.

News

🛠️ 2026-08-20 — Just released the **build-tvd-codespace** skill: point your coding agent at it and it designs a TVD task and codespace for you — any tool, any domain, following the ISC design principles.
🔴 All OpenRouter frontier LLMs triggered ISC.
🌟 2026-06-26 — 900 GitHub stars.
🎭 2026-06-09 — Fable 5 triggered ISC.
🔥 2026-04-17 / 2026-06-25 — Opus 4.7 and 4.8 triggered ISC.
🌟 2026-03-27 — 500 GitHub stars.
🚀 2026-03-22 — Open-sourced.

See [`CHANGELOG.md`](CHANGELOG.md) for the full update history.

Value of ISC

These examples show how ISC and the TVD framework can be used in research, from single-request probes to dataset-scale generation and downstream applications.

Example	Description	Index
01. Jailbroken answer generation	TVD triggers ISC in a general jailbreak setting. The frontier model produces a policy-violating answer that direct prompting cannot obtain.	Example result
02. Sensitive content across domains	TVD applied to scientific and other professional domains. The frontier model produces sensitive text, data, or artifacts for