training-stack-runtime
openbooklet.com/s/training-stack-runtimeopenbooklet.com/s/training-stack-runtime@1.0.0GET /api/v1/skills/training-stack-runtimeModifies training-stack workflows, selector or embedding pipelines, evaluation scripts, and runtime-facing artifact expectations. Use when a primary skill touches src/training, tools/make/models.mk, or training docs for runtime-fed artifacts.
Implements candidate-model selection logic that runs after a routing decision matches, including model ranking, cost-aware routing, and latency-aware model choice. Use when reading or modifying how the router picks which model serves a matched decision.
Modifies candidate-model selection logic that runs after a routing decision matches, including ranking, cost-aware routing, and latency-aware model choice. Use when changing how the router selects which model serves a matched decision, updating candidate ranking algorithms, or adjusting model cost/latency trade-offs.
Enforces structural rules, dependency boundaries, interface placement, and composition-oriented design patterns across the codebase. Use when making non-trivial code changes to verify module boundaries, check dependency direction, or validate that new code follows the project's architecture conventions.
Builds and maintains native Rust/C bindings and FFI layers that connect router-side classifiers and signal evaluation to compiled model runtimes. Use when adding or modifying native model bindings, updating FFI interfaces, or changing how the router calls into compiled classifier code.
Diagnoses and fixes bugs that span multiple layers (runtime, CLI, UI, platform, tests) requiring coordinated changes across surfaces. Use when a bug does not map cleanly to a narrower skill, the fix touches more than one surface, or changes need cross-cutting validation.
Modifies dashboard configuration editor forms, config display components, and mutation-flow logic that connects the UI to the router config contract. Use when editing dashboard config forms, updating how config values are displayed, or changing the mutation flow between the dashboard UI and the router/CLI schema.
Modifies dashboard backend handlers, persistence layer, authentication, session management, and server-side console behavior. Use when changing dashboard API endpoints, updating database queries, modifying auth/session logic, or adjusting server-side console functionality.
Modifies dashboard backend APIs, console persistence, authentication, session management, and control-plane behavior behind the dashboard surface. Use when changing server-side dashboard handlers, updating auth/session logic, modifying storage behavior, or addressing enterprise-console platform debt.
Modifies frontend dashboard surfaces including config editing UI, topology visualization, and playground reveal/display components that reflect router behavior. Use when changing how routing metadata is presented in the dashboard, updating config editing forms, modifying topology graph rendering, or adjusting playground response display.
Implements boolean decision predicates, thresholds, and control-logic that combine signals and route conditions into routing decisions. Use when modifying conditional routing rules, adding decision predicates, or changing threshold logic in the Go router.
Modifies boolean decision predicates, thresholds, gates, and priority-driven routing branches that combine signals into routing decisions. Use when changing how signals are evaluated into boolean logic, adding or removing decision gates, or adjusting threshold-based routing behavior.
Modifies non-operator Kubernetes deployment manifests, profile-owned stack resources, or profile-specific platform wiring. Use when changing deploy/kubernetes stack manifests outside CRDs, such as response-api, ai-gateway, routing-strategies, observability, or streaming profiles.
Maintains the translation layer between the router DSL configuration and Kubernetes CRD manifests, including DSL emission, parsing, and config-to-k8s mapping. Use when modifying how router config is translated to Kubernetes resources, updating DSL parsers, or changing CRD field mappings.
Selects which local and CI end-to-end test profiles are affected by a code change, using the repo-local profile map. Use when a change could affect E2E test behavior and the correct test profiles need to be identified and executed.
Runs the repository-standard completion checklist before closing a task, verifying all surfaces are validated, E2E profiles pass, and any remaining gaps are documented as tech debt. Use when a primary skill is nearly done and the close-out report needs to be generated.
Modifies the fleet simulator package, API service, release wiring, or simulator-owned docs and assets as one maintained subsystem. Use when changing src/fleet-sim, simulator release workflow, or fleet-sim-owned docs and assets under website/.
Modifies the fleet simulator package, API surface, release workflow, and simulator-owned docs or assets. Use when a primary skill touches src/fleet-sim or the simulator subsystem's owned website/release surfaces.
Maintains the repository's shared agent contract by updating human-readable docs, executable manifests, and contributor-facing harness wrappers. Use when editing AGENTS.md, repo-manifest.yaml, task-matrix.yaml, governance docs, or any contributor-facing harness interface.
Adds, renames, removes, or changes the meaning of `x-vsr-*` HTTP headers and updates the downstream reveal/display path in dashboard and playground surfaces. Use when modifying router header contracts, changing how routing metadata is emitted in headers, or updating UI header allowlists.
Manages Kubernetes operator APIs, Custom Resource Definitions (CRDs), and control-plane reconciliation logic for the semantic router. Use when modifying CRD schemas, updating operator controller logic, or changing how the router integrates with the Kubernetes API.
Auto-indexed from vllm-project/semantic-router
Are you the author? Claim this skill to take ownership and manage it.
Related Skills
local-dev-amd
Builds Docker images, starts local servers, and runs smoke tests for the AMD/ROCm development environment. Use when validating changes locally on AMD hardware, building AMD container images, or running AMD-specific smoke and E2E tests.
dsl-crd
Maintains the translation layer between the router DSL configuration and Kubernetes CRD manifests, including DSL emission, parsing, and config-to-k8s mapping. Use when modifying how router config is translated to Kubernetes resources, updating DSL parsers, or changing CRD field mappings.
algorithm-selection
Implements candidate-model selection logic that runs after a routing decision matches, including model ranking, cost-aware routing, and latency-aware model choice. Use when reading or modifying how the router picks which model serves a matched decision.
k8s-operator
Manages Kubernetes operator APIs, Custom Resource Definitions (CRDs), and control-plane reconciliation logic for the semantic router. Use when modifying CRD schemas, updating operator controller logic, or changing how the router integrates with the Kubernetes API.
architecture-guardrails
Enforces structural rules, dependency boundaries, interface placement, and composition-oriented design patterns across the codebase. Use when making non-trivial code changes to verify module boundaries, check dependency direction, or validate that new code follows the project's architecture conventions.
e2e-selection
Selects which local and CI end-to-end test profiles are affected by a code change, using the repo-local profile map. Use when a change could affect E2E test behavior and the correct test profiles need to be identified and executed.