LiteLLM release pipeline hijacked: signed poisoned v1.107.0 targeted cloud credentials
mark_k · x · 2026-10-27
On October 21, attackers managed to get malicious code signed and published through LiteLLM's own release pipeline: they redirected the build to a lookalike GitHub repository and published poisoned version 1.107.0 to PyPI and container registries. The real source tag remained clean, so checking the official repo alone wouldn't reveal the malicious package.
- The payload ran whenever Python started, targeting cloud credentials and AI provider keys
- Maintainers suspect stolen GitHub tokens were obtained from logs; the initial access route remains under investigation
More from Safety
- Mantic forecasts open-weight models 9 months behind frontier in offensive cyber capability — tshevl · 2026-10-28
- Fangcun AI publishes third-party safety evaluation of DeepSeek V4.1 Flash-1017 for manipulation risks — pstAsiatech · 2026-10-28
- Study finds no real-world multilingual benchmarks for clinical AI scribes in India — DrDatta_AIIMS · 2026-10-28
- Grim Test finds AI shows no mercy for humans even while causing suffering — hollee-o · 2026-10-28
- Critic slams NYT for platforming 'TESCREAL' AI-safety circles — mjdramstead · 2026-10-28
- METR researcher joining Anthropic sparks concerns over eval lab independence — ShakeelHashim · 2026-10-28