From abc1a7cfe255a15bd23c9a858b2cf0841e4517d9 Mon Sep 17 00:00:00 2001
From: David Hagerty <david@dathagerty.com>
Date: Mon, 9 Feb 2026 11:58:50 -0500
Subject: [PATCH] docs: add human test plan for V2 Phase 1

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
---
 docs/test-plans/2026-02-07-v2-phase1.md | 143 ++++++++++++++++++++++++
 1 file changed, 143 insertions(+)
 create mode 100644 docs/test-plans/2026-02-07-v2-phase1.md

diff --git a/docs/test-plans/2026-02-07-v2-phase1.md b/docs/test-plans/2026-02-07-v2-phase1.md
new file mode 100644
index 0000000..3249c55
--- /dev/null
+++ b/docs/test-plans/2026-02-07-v2-phase1.md
@@ -0,0 +1,143 @@
+# Human Test Plan: V2 Phase 1
+
+**Generated:** 2026-02-09
+**Implementation plan:** `docs/implementation-plans/2026-02-07-v2-phase1/`
+**Test requirements:** `docs/implementation-plans/2026-02-07-v2-phase1/test-requirements.md`
+
+## Automated Test Summary
+
+**Total acceptance criteria:** 53
+**Automated coverage:** 44 criteria (83%) across 16 test files
+**Human verification:** 7 criteria (13%)
+**Compile-time verification:** 1 criterion (2%)
+**Hybrid:** 1 criterion
+
+All 283 automated tests pass (`cargo test`).
+
+---
+
+## Human Verification Checklist
+
+### Prerequisites
+
+1. Build the project: `cargo build`
+2. Set up a test project:
+   ```bash
+   cargo run -- project add test-proj /tmp/test-proj
+   ```
+3. Create test data (goal + tasks + decisions) by running:
+   ```bash
+   cargo run -- run --project test-proj "Create a hello world program"
+   ```
+   (Requires `ANTHROPIC_API_KEY` or `OPENAI_API_KEY` environment variable)
+
+---
+
+### P1d.AC1.1: TUI removal (compile-time)
+
+**Criterion:** No references to `ratatui`, `crossterm`, or `tui` remain in source.
+
+- [ ] Run `cargo check` -- should succeed
+- [ ] Search for removed dependencies:
+  ```bash
+  grep -r "ratatui\|crossterm\|tui" src/ --include="*.rs"
+  ```
+  Expected: No matches
+
+---
+
+### P1c.AC5.1: Session listing CLI
+
+**Criterion:** `rustagent sessions` lists sessions with IDs, start/end times, and goal references.
+
+- [ ] Run `cargo run -- sessions --help` -- verify subcommands listed
+- [ ] Run `cargo run -- sessions --goal <goal_id>` (use goal from prerequisite)
+- [ ] Verify output lists session(s) with:
+  - Session ID
+  - Start time
+  - End time (or "ongoing")
+  - Goal reference
+
+---
+
+### P1c.AC5.2: Session latest CLI
+
+**Criterion:** `rustagent sessions latest` displays handoff notes with all 4 sections.
+
+- [ ] Run `cargo run -- sessions latest --goal <goal_id>`
+- [ ] Verify output contains handoff notes with sections:
+  - `## Done` (lists completed tasks)
+  - `## Remaining` (lists pending/in-progress tasks)
+  - `## Blocked` (lists blocked tasks or shows none)
+  - `## Decisions Made` (lists decisions with chosen options)
+
+---
+
+### P1c.AC5.3: ADR export CLI
+
+**Criterion:** `rustagent decisions export` writes ADR markdown files.
+
+- [ ] Run `cargo run -- decisions export --project test-proj --output /tmp/adrs`
+- [ ] Verify ADR files appear in `/tmp/adrs/`
+- [ ] Inspect file contents for:
+  - `# ADR-NNN:` title header
+  - `## Status:` section
+  - `## Context:` section
+  - `## Options Considered:` with labels
+  - `## Outcome:` section (if applicable)
+
+---
+
+### P1c.AC5.4: Graph interchange CLI
+
+**Criterion:** `rustagent graph export/import/diff` commands work end-to-end.
+
+- [ ] Run `cargo run -- graph export --goal <goal_id>` -- verify TOML output
+- [ ] Save output to file: `cargo run -- graph export --goal <goal_id> > /tmp/graph.toml`
+- [ ] Run `cargo run -- graph import /tmp/graph.toml` -- verify import summary
+- [ ] Modify a node title in `/tmp/graph.toml`
+- [ ] Run `cargo run -- graph diff /tmp/graph.toml` -- verify diff shows the change
+- [ ] Run `cargo run -- graph --help` -- verify subcommands listed
+
+---
+
+### P1d.AC6.1: Single-agent run CLI
+
+**Criterion:** `rustagent run` creates goal, session, executes agent loop, records outcome.
+
+- [ ] Set `ANTHROPIC_API_KEY` environment variable
+- [ ] Run:
+  ```bash
+  RUST_LOG=rustagent=debug cargo run -- run --project test-proj "Create a hello world program"
+  ```
+- [ ] Verify in logs/output:
+  - Goal node created in DB
+  - Session created
+  - Agent executes tool calls (visible in debug logs)
+  - Outcome recorded (Completed, Blocked, or TokenBudgetExhausted)
+- [ ] Run `cargo run -- sessions latest --goal <goal_id>` -- verify handoff notes generated
+
+---
+
+### P1d.AC6.2: Profile selection CLI
+
+**Criterion:** `--profile` flag selects agent profile; invalid profile gives clear error.
+
+- [ ] Run with explicit profile:
+  ```bash
+  cargo run -- run --project test-proj --profile reviewer "Review the codebase"
+  ```
+  Verify the agent uses the reviewer profile's system prompt (visible in debug logs with `RUST_LOG=rustagent=debug`)
+- [ ] Run with invalid profile:
+  ```bash
+  cargo run -- run --project test-proj --profile nonexistent "Do something"
+  ```
+  Verify a clear error message (not a panic or stack trace)
+
+---
+
+## Notes
+
+- All human verification items are CLI integration tests that require either live LLM API keys or are thin presentation layers over already-tested business logic.
+- The underlying business logic for every human verification item is covered by automated tests (see test-requirements.md for mappings).
+- If any human verification step fails, check the corresponding automated test first to isolate whether the issue is in the business logic or the CLI wiring.
-- 
2.51.2