Commit History

19 days ago

  1. fix: sync proxy stream inspection, canary matching, and layer validation to HF Space Running

    c7b149d
  2. sec: enforce .env ignores and update auth middleware

    344b3b7

  1. fix(mongo): increase connection timeout to 15s with retries

    d450618
  2. fix: allow system environment variables to take priority in main.py

    60105e7
  3. sync: push updated backend files (auth fix and cleanups)

    860cdea
  4. fix(auth): support X-Forwarded-For client IP and bump rate limit

    3d913a7

  1. chore: prune dead weight, remove unused dependencies and pycache

    384aa25

  1. chore: sync updated backend to HF

    8331c7e

  1. chore: sync updated backend to HF

    36e4dac

  1. chore: sync updated backend to HF

    85c6c01
  2. SEC: Fix critical multi-tenant data leak in Threat Graph

    928ec36
  3. Feature: Sankey-style three-stage attack pipeline (Key->Attack->Layer)

    e21ce30
  4. API: Return API Key names instead of IDs in graph stats

    4aa30d5
  5. UI: Improve force graph collisions and velocity chart window

    8cd8d5d
  6. Implement Threat Intelligence dashboard and capture session/client identity

    cd91aab
  7. Fix attack type normalization and safe classification leak in graph

    08e7bbd
  8. Architectural optimizations and frontend fixes

    d468fc4

  1. Integrate Neo4j Threat Graph & Dashboard

    814db53

  1. fix: match pepper fallback to local env so existing keys work

    aba06d0
  2. fix: add fallback for API_KEY_PEPPER

    a5fce17
  3. chore: complete ui refactor and security enhancements

    90cbfa2

  1. Configure 6-class model and deploy to HF Spaces

    0d4a57c

  1. feat(ui): dynamically pull active layer count from backend health endpoint

    78ca80a
  2. fix(ui): update Logs page filters to use real attack vectors and tripped layers instead of outdated ingress nodes

    5a904c0
  3. feat(security): disable OpenAI moderation layer completely per user request

    3a62ceb
  4. fix(security): add PROMPT_EXTRACTION regex rules to catch explicit system prompt leaks

    aa8c9a7
  5. feat(security): transition firewall to a cumulative risk ensemble architecture

    6a3e3cc
  6. fix(security): enforce pipeline risk threshold over openai moderation boolean

    dee921b
  7. fix(ui): add openai moderation to live monitor breakdown and fix 0 score parsing

    363f7ca
  8. feat(security): per-api-key opt-in for openai moderation layer

    17d0fd8
  9. feat(backend): add optional openai moderation layer

    19eb0be

  1. v1.0.0

    f3fa932