Researchers pull real leaked API keys from LLMs using only API output
nordicinst · x · 2026-10-20
A GitGuardian blog post details how researchers recovered real, leaked DeepSeek and Google API keys from Qwen3-27B, GPT-Codex 5.3, and Claude Haiku 4.5 using only API output access — every key had a matching public GitHub occurrence.
How agents pull secrets in: credentials live in .env files, shell history, and MCP configs; coding agents reading configs, running commands, or calling MCP servers place them into model context. Claude Code-assisted commits leaked secrets at 3.2% versus a 1.5% baseline.
Mitigations: block secrets before the model sees them — AI Hooks intercept secrets in prompts, tool calls, and output, while Developer Endpoint Protection reduces plaintext credentials on managed fleets.
More from coding & agent
- Dev builds Overflow to turn leftover LLM subscription usage into open-source fixes — Nitjsefnie · 2026-10-20
- mlx-geospatial launches: fast geospatial Python library on Apple silicon via MLX — HankYeomans · 2026-10-20
- Mermaid's new Pele update brings major speedups and real-time color themes — draginol · 2026-10-20
- Vibe coding is everywhere, but new products aren't — discipline is the bottleneck — mariolefebvre · 2026-10-20
- RL fine-tuning Qwen3.5-0.8B for under $20 on HF Jobs lifts pass@1 from 10% to 36% — LysandreJik · 2026-10-20
- DeepSeek beats Claude and Codex on Google Maps scraping speed, at a steep price — teortaxesTex · 2026-10-20