Open-source AI teammates in Slack, Teams, Discord, Telegram and Mattermost. Self-host it, with any agent harness and any model.
Python agent-orchestrationaiai-agentschatopsdiscordlangchainmcpmicrosoft-teamsmulti-agentopen-sourceopenaiself-hostedslack
Document thread replies during an outage and the Strands image's licences
agent-runner353fd1c61dcd0ff
diff --git a/console/packages/plugins/src/switch-skill/SKILL.md b/console/packages/plugins/src/switch-skill/SKILL.mdindex 543699da7..e524fb786 100644--- a/console/packages/plugins/src/switch-skill/SKILL.md+++ b/console/packages/plugins/src/switch-skill/SKILL.md@@ -155,9 +155,11 @@ does, and there is no polling tool to call.pick wrong: if the message is not really yours to answer (people talking toeach other, or something another agent covers), stay silent and postnothing. In such a room, replies in a thread go to the decision service too,- rather than to the agent that spoke there last, and answering a routed- message does not make the person's next untagged message a follow-up to you:- it is decided again. `!auto-route off` turns it off, and `!auto-route` on its+ rather than to the agent that spoke there last (so while the decision+ service is down, a reply in your thread that doesn't tag you reaches you+ only if you are the room's triage agent, and otherwise nobody), and+ answering a routed message does not make the person's next untagged message+ a follow-up to you: it is decided again. `!auto-route off` turns it off, and `!auto-route` on itsown says which it is and whether it can act. Anyone who can post in theroom may turn it on or off, from the room or from its toggle in SwitchConsole or on the gateway's room page (both need write access to the room).diff --git a/console/packages/plugins/src/switch-skill/switch-skill.test.ts b/console/packages/plugins/src/switch-skill/switch-skill.test.tsindex e77958abb..15c9a42a6 100644--- a/console/packages/plugins/src/switch-skill/switch-skill.test.ts+++ b/console/packages/plugins/src/switch-skill/switch-skill.test.ts@@ -29,6 +29,13 @@ describe('switch skill', () => {expect(text).toContain("came to you as the room's triage agent");});+ it('says where an untagged thread reply goes while the decision service is down', () => {+ const text = SWITCH_SKILL_CONTEXT.replace(/\s+/g, ' ');+ expect(text).toContain(+ "while the decision service is down, a reply in your thread that doesn't tag you reaches you only if you are the room's triage agent, and otherwise nobody"+ );+ });+it('says who may turn auto-route on or off, and who may name the triage agent', () => {const text = SWITCH_SKILL_CONTEXT.replace(/\s+/g, ' ');expect(text).toContain('Anyone who can post in the room may turn it on or off');diff --git a/decision/README.md b/decision/README.mdindex df186f4e3..769d5f973 100644--- a/decision/README.md+++ b/decision/README.md@@ -183,6 +183,33 @@ To use Haiku only, no model server is needed: set`DECISION_BACKEND=anthropic` and `ANTHROPIC_API_KEY`, or leave the defaultbackend and send the override headers per request.+## Licences of the Strands image++`switch-decision-strands` installs `strands-decider` (Apache-2.0) and every+dependency pinned by hash in `model-servers/strands/requirements.txt`. Most+are under permissive licences (Apache-2.0, MIT, BSD). Two groups are worth+naming, and licence scanners flag the second:++- **PyTorch** (`torch`): BSD-3-Clause.+- **NVIDIA's CUDA runtime libraries**, the `nvidia-*` wheels PyTorch depends+ on (cuBLAS, cuDNN, cuFFT, cuFile, cuRAND, cuSOLVER, cuSPARSE, NVRTC, CUPTI,+ NVTX, NCCL and the rest): proprietary, under NVIDIA's end user licence+ agreement for CUDA, which allows them to be redistributed as part of an+ application that uses them. Their package metadata says "Other/Proprietary+ License", which is what the scanners report. The base image,+ `nvidia/cuda:13.0.3-base-ubuntu24.04`, comes under NVIDIA's terms too.++**The decision: ship them.** The model server needs CUDA to run on a GPU, and+these are NVIDIA's own runtime packages, used unmodified and in the way the+EULA provides for. Nothing of them is in Switch's source: this repository+names them, pinned by version and hash, and nothing more. The image is built+by Switch's release CI and published to its registry, and a deployment pulls+it at deploy time, in the same way it pulls the model weights from Hugging+Face; neither is vendored into the source. Anyone who can't accept NVIDIA's+terms can run without this image: Qwen on llama.cpp's CPU image (compose's+`decision` profile), or Haiku with no model server+(`decision.modelServer.enabled: false` in the chart).+## Evaluate`eval/` holds two scenario sets: `scenarios.json` (15 easy cases, three agents)diff --git a/docs/old/decision-routing.md b/docs/old/decision-routing.mdindex 4c1160271..f591e7289 100644--- a/docs/old/decision-routing.md+++ b/docs/old/decision-routing.md@@ -156,6 +156,13 @@ a failure. A decision of nobody is not.so it knows why it got it. It cannot be the room's assigned agent, whichalready gets every untagged message before auto-route is asked.`!auto-route fallback off` clears it.+- **Thread replies too.** In a room with auto-route on, a reply in a thread+ is decided like any other untagged message: the follow-up rule that sends+ it to the agent that spoke there last is skipped. So while the backend is+ failing, a reply in an agent's thread that doesn't tag it goes to the+ room's triage agent if one is named, and otherwise to nobody, with the+ notice above; in a room without auto-route it would have reached that+ agent. People reach it meanwhile by tagging it, which works as always.### Across a restart