vac-testlisted

# VAC Test Generate test scenarios, evaluate transcripts, iterate. Output: `test-scenarios.md` + `test-results.md`. ## Requirements (input check) Before starting, check that these files exist in the agent directory (`../voice-agents/<agent>/`): | File | Comes from | Missing? | |---|---|---| | `briefing.md` | `vac-intake` (step 1) | Run `vac-intake` first | | `conversation-design.md` | `vac-design` (step 3) | Run `vac-design` first | | `prompt.md` | `vac-prompt` (step 4) | Run `vac-prompt` first | | `AGENT.md` (with `agent_id`) | `vac-deploy` (step 6) | Run `vac-deploy` first; no live deployed agent means no test call | | `deployment-log.md` | `vac-deploy` (step 6) | Run `vac-deploy` first | If an input is missing, do NOT improvise and do NOT carry on with placeholders: name the missing upstream step and ask the client or the workspace admin whether it should be made up now. ## Process 1. Read all artifacts (briefing through deployment) 2. Generate test scenarios covering all conversation paths 3. **WAIT for human test calls** — provide scenarios to the test callers named by the requester (the client) 4. Receive and evaluate transcripts 5. Score each call, identify patterns 6. Feed findings back into prompt (return to vac-prompt if needed) 7. Max 3 test-iterate cycles, then finalize ## Test Scenario Generation ### Coverage Requirements Every test suite MUST cover: | Category | Min Scenarios | Description | |---|---|---| | **Happy Path** | 2 | Perfect conversation, p