security, inference, and whatever rabbit hole i'm in this week.
Pinned Loading
-
zyphr-labs/bounty-harness
zyphr-labs/bounty-harness PublicAgent harness for authorized bug bounty research with safety tiers, autonomous campaigns, reusable security skills, and evidence-first reporting
Python 1
-
confidence-tax-study
confidence-tax-study PublicDoes medical-reasoning fine-tuning make a small LLM safer or confidently wrong? A controlled before/after calibration study on Qwen3-4B + MedGemma-4B (single free Kaggle T4). Finding: internal cali…
Python
-
zyphr-labs/pii-guard
zyphr-labs/pii-guard PublicBrowser extension + local proxy that strips PII from LLM requests before they leave your machine. Presidio + spaCy NER, all detection runs locally.
JavaScript 1
-
dsh-medseek
dsh-medseek PublicClinical tools for DeepSeek Harness: handover and discharge drafts, de-identification, completeness checks, and cited lookups. Draft-only, not a medical device.
TypeScript 2
If the problem persists, check the GitHub status page or contact support.




