---
phase: 2
slug: agent-features
status: draft
nyquist_compliant: false
wave_0_complete: false
created: 2026-03-23
---

# Phase 2 — Validation Strategy

> Per-phase validation contract for feedback sampling during execution.

---

## Test Infrastructure

| Property | Value |
|----------|-------|
| **Framework** | pytest 8.x + pytest-asyncio (existing from Phase 1) |
| **Config file** | `pyproject.toml` (existing `[tool.pytest.ini_options]`) |
| **Quick run command** | `pytest tests/unit -x -q` |
| **Full suite command** | `pytest tests/ -x` |
| **Estimated runtime** | ~45 seconds |

---

## Sampling Rate

- **After every task commit:** Run `pytest tests/unit -x -q`
- **After every plan wave:** Run `pytest tests/ -x`
- **Before `/gsd:verify-work`:** Full suite must be green
- **Max feedback latency:** 45 seconds

---

## Per-Task Verification Map

| Task ID | Plan | Wave | Requirement | Test Type | Automated Command | File Exists | Status |
|---------|------|------|-------------|-----------|-------------------|-------------|--------|
| 02-01-xx | 01 | 1 | AGNT-02 | unit | `pytest tests/unit/test_memory_short_term.py -x` | ❌ W0 | ⬜ pending |
| 02-01-xx | 01 | 1 | AGNT-03 | integration | `pytest tests/integration/test_memory_long_term.py -x` | ❌ W0 | ⬜ pending |
| 02-02-xx | 02 | 1 | AGNT-04 | unit | `pytest tests/unit/test_tool_registry.py tests/unit/test_tool_executor.py -x` | ❌ W0 | ⬜ pending |
| 02-02-xx | 02 | 1 | AGNT-06 | integration | `pytest tests/integration/test_audit.py -x` | ❌ W0 | ⬜ pending |
| 02-03-xx | 03 | 2 | CHAN-03 | unit | `pytest tests/unit/test_whatsapp_verify.py tests/unit/test_whatsapp_normalize.py -x` | ❌ W0 | ⬜ pending |
| 02-03-xx | 03 | 2 | CHAN-04 | unit | `pytest tests/unit/test_whatsapp_scoping.py -x` | ❌ W0 | ⬜ pending |
| 02-04-xx | 04 | 2 | AGNT-05 | unit+integ | `pytest tests/unit/test_escalation.py tests/integration/test_escalation.py -x` | ❌ W0 | ⬜ pending |

*Status: ⬜ pending · ✅ green · ❌ red · ⚠️ flaky*

---

## Wave 0 Requirements

- [ ] `tests/unit/test_memory_short_term.py` — AGNT-02: Redis sliding window
- [ ] `tests/integration/test_memory_long_term.py` — AGNT-03: pgvector retrieval + tenant isolation
- [ ] `tests/unit/test_tool_registry.py` — AGNT-04: tool registry lookup
- [ ] `tests/unit/test_tool_executor.py` — AGNT-04: schema validation + confirmation
- [ ] `tests/integration/test_audit.py` — AGNT-06: audit immutability
- [ ] `tests/unit/test_escalation.py` — AGNT-05: transcript packaging
- [ ] `tests/integration/test_escalation.py` — AGNT-05: DM delivery
- [ ] `tests/unit/test_whatsapp_verify.py` — CHAN-03: webhook signature verification
- [ ] `tests/unit/test_whatsapp_normalize.py` — CHAN-03: message normalization
- [ ] `tests/unit/test_whatsapp_scoping.py` — CHAN-04: business-function gate
- [ ] `tests/conftest.py` — extend with pgvector fixtures, mock MinIO (moto)
- [ ] Install: `uv add --dev moto` (S3/MinIO mocking)

---

## Manual-Only Verifications

| Behavior | Requirement | Why Manual | Test Instructions |
|----------|-------------|------------|-------------------|
| WhatsApp message delivery via live Cloud API | CHAN-03 | Requires Meta-verified WABA + phone number | Send test message to configured WhatsApp number, verify reply |
| Media (image) round-trip via WhatsApp | CHAN-03 | Requires live WhatsApp + multimodal LLM | Send photo to agent, verify interpretation; request doc from agent |

---

## Validation Sign-Off

- [ ] All tasks have `<automated>` verify or Wave 0 dependencies
- [ ] Sampling continuity: no 3 consecutive tasks without automated verify
- [ ] Wave 0 covers all MISSING references
- [ ] No watch-mode flags
- [ ] Feedback latency < 45s
- [ ] `nyquist_compliant: true` set in frontmatter

**Approval:** pending