kfchow-llm-value-router.yaml 2.2 KB

123456789101112131415161718192021222324252627282930313233343536
  1. # Community catalog entry — kfchow-llm-value-router
  2. #
  3. # Owner-submitted: the PR author (kfchow-ai) owns and maintains the repo below.
  4. # Exact-SHA pin (40-hex, resolves as a COMMIT on the repo). No self-updating code: the resolver fetches a public
  5. # data feed (kfchow.com value leaderboard), never the plugin's own files.
  6. name: kfchow-llm-value-router
  7. repo: https://github.com/kfchow-ai/kfchow-llm-value-router
  8. sha: "39a6944280c2156f816df8641d3d01d743aec564"
  9. description: >-
  10. Per-turn LLM routing: on each matching turn a TypeSafe Jev classifier call (api.typesafe.ai) receives turn-shape
  11. features plus the first 1200 chars of the last user message (shadow mode included) and labels it free/paid;
  12. high-confidence routine turns are downgraded to the provider's configured free rung, strategic mid-rung turns on
  13. Nous are escalated to the configured premium rung after a 2-token billed credit probe (Nous only, fail-closed).
  14. Fail-open, shadow-by-default, per-provider rungs from the user's lane2_config.json (leaderboard resolver script
  15. optional). Probe model id is fixed in code (z-ai/glm-5.3-flash). Disclosure — once you list models in
  16. ~/.hermes/jev/lane2_config.json, the first call of each matching turn sends message/tool counts and the first 1,200
  17. characters of the last user message (send_excerpt, on by default) to TypeSafe Jev at api.typesafe.ai with your
  18. JEVI_API_KEY, in shadow mode as well as live; in live mode Jev's answer switches the request (and every later call
  19. of that turn, via an llm_execution wrapper) to the free, flash or premium model you configured on the same provider;
  20. with escalation set up on Nous it sends a 2-token billed ping to the Nous inference API with NOUS_API_KEY at most
  21. every 5 minutes; every turn appends a decision row (with that excerpt on routed turns) to
  22. ~/.hermes/jev/lane2-live.jsonl, which is never rotated.
  23. maintainer: KFChow
  24. tier: community
  25. category: models
  26. requires_hermes: ">=0.21.5"
  27. version: "1.0.6"
  28. platforms: [linux, macos]
  29. docs_url: https://github.com/kfchow-ai/kfchow-llm-value-router
  30. capabilities:
  31. provides_tools: []
  32. provides_hooks: []
  33. provides_middleware:
  34. - llm_request
  35. - llm_execution
  36. requires_env: []