OpenParable
Skip to content
← Back to skills

Signal End To End

Development
signal-end-to-end

Signal End To End is an Agent Skill for development published in the vllm-project/semantic-router repository on GitHub: a SKILL.md instruction file an AI agent can load to do this task.

Implements end-to-end signal changes spanning router config, signal extraction, CLI schema, optional bindings, router-owned metadata headers, and E2E test coverage. Use when adding a new signal type, changing signal configuration or extraction logic, updating CLI schema for signal parameters, or modifying router-owned signal metadata contracts.

View on GitHub ↗
QUICK START

How to use this skill

Bring this guide into your coding agent with a prompt tailored to the tool you use.

  1. Open your project in Codex.
  2. Copy and paste the prompt below.
  3. Review files and risks before approving any action.
Prompt to paste
Inspect the Agent Skill at https://github.com/vllm-project/semantic-router/blob/HEAD/tools/agent/skills/signal-end-to-end/SKILL.md and all supporting files. Report provenance, permissions, scripts, dependencies, and security risks. Propose an exact installation into .agents/skills/signal-end-to-end, but wait for my approval before writing files or running scripts.

Copying this prompt does not install or run the skill. Review third-party files before use.

Signal End to End

Trigger

  • Add a new signal type or signal rule shape
  • Change how an existing signal is configured, extracted, emitted, or displayed
  • Touch router signal config plus Python CLI schema in the same feature
  • Change router-owned routing metadata headers or topology payloads that are derived from signal behavior

Workflow

  1. Read change surfaces and module boundaries for signal context
  2. Modify signal config, extraction, CLI schema, and dashboard surfaces as needed
  3. Run make agent-report ENV=cpu CHANGED_FILES="..." to identify all impacted surfaces
  4. Run make agent-ci-gate CHANGED_FILES="..." to validate signal contract alignment
  5. When the signal becomes user-visible, verify the router-owned metadata chain stays aligned across classify or eval payloads, emitted headers or looper headers, and any dependent reveal or topology consumers
  6. Verify router config, signal extraction, and CLI schema stay aligned with relevant E2E coverage

Gotchas

  • Signal changes rarely stop at extraction; CLI schema, router-owned metadata headers, or bindings often need updates too.
  • New matched-signal families often need more than one UI touchpoint: matched_signals JSON, x-vsr-matched-* headers, chatRequestSupport, HeaderDisplay, HeaderReveal, topology backend conversion, and topology frontend parser or constants.
  • New signal semantics need explicit E2E ownership, not just unit-level classifier coverage.
  • If the header or topology contract originates from signal evaluation, keep it in this primary instead of inventing a header-only flow.

Must Read

Standard Commands

  • make agent-report ENV=cpu CHANGED_FILES="..."
  • make agent-ci-gate CHANGED_FILES="..."
  • make agent-feature-gate ENV=cpu CHANGED_FILES="..."

Acceptance

  • Router config, signal extraction, and Python CLI schema stay aligned for the signal contract
  • Any router-owned signal metadata updates the relevant classify or eval payloads, emitted headers, and dependent reveal or topology consumers together
  • Relevant E2E coverage is added or updated when behavior changes