Labels
Labels
area/auth
area/billing
area/compression
area/config
area/i18n
area/install-update
area/local-models
Local model inference/runtimes: llama.cpp, Ollama, LM Studio, MLX/vLLM, GGUF, VRAM and offloadarea/profiles
area/sessions
area/streaming
awaiting-reporter
Parked: needs info/experiment from the reporter before any work; revisit only when they replybackend/daytona
backend/docker
backend/local
backend/modal
backend/other
A terminal execution backend without its own label (for example Singularity/Apptainer, Vercel Sandbobackend/ssh
blocked
bug
cause/fallout
cause/upstream
A cause outside Hermes that Hermes cannot fully fix: a model provider's outage, policy change, rateci-reviewed
comp/acp
comp/agent
Core agent runtime: loop, agent_init, prompt builder, context-compression, responses endpointcomp/cli
comp/cron
comp/dashboard
comp/desktop
comp/gateway
comp/plugins