OpenCode and LiteLLM breached in one week, both attacks hunted AI API keys
victorialslocum · x · 2026-10-22
Two supply-chain security incidents in one week both targeted AI API keys:
- OpenCode: an intruder entered the legacy Console system (slated for retirement) on the night of Oct 16–17, exposing stored AI keys for about 1.4% of console users.
- LiteLLM: a poisoned v1.107.0 was published to PyPI on Oct 21 via stolen GitHub credentials; on Python start it could steal SSH keys, cloud credentials, AI tool logins, and every stored provider key on gateway servers, sending them to a lookalike domain.
The author notes that after OpenCode, common advice was to self-host a gateway — yet LiteLLM itself is now a target, because anything holding all your API keys is the highest-value attack surface. Three takeaways: keep API keys in as few places as possible, set budget caps on every AI provider, and enforce least privilege everywhere so a stolen credential can't reach everything.
More from coding & agent
- swyx hails Opus 5.5 as best work-estimator model; dev counters that review time is the real bottleneck — petrusenko_max · 2026-10-23
- Agent Field Kit launches free agentic defense toolkit against LLM-enhanced attackers — alvelda · 2026-10-23
- LiteLLM PyPI compromise and Railway breach: Oct 21 attack timeline and free response kit — alvelda · 2026-10-23
- Daily Experiment Planner agent: one prompt to save scientists hours of lab prep — nlarusstone · 2026-10-23
- Two £300 Snapdragon mini PCs, no GPU: 46 tok/s local Qwen3.8 Flash Next with custom ToRc engine — WorldlyAd7946 · 2026-10-22
- Chrome engineer asks: do agents improve incident response or just ship unpatched deps 10x faster? — fhinkel · 2026-10-22