vtx-coding-agent 0.1.1__tar.gz

vtx_coding_agent-0.1.1/.agents/skills/vtx-release-publish/SKILL.md

@@ -0,0 +1,98 @@
+---
+name: vtx-release-publish
+description: Tag, publish to PyPI, and create GitHub release for Vtx with validation and rollback-safe steps
+---
+
+# Vtx Release + PyPI Publish
+
+Use this skill when the user asks to cut a new Vtx version, tag it, publish to PyPI, and/or create a GitHub release.
+
+## Inputs to confirm
+
+- Target version (example: `0.2.1`)
+- Base range for notes (usually previous tag, example: `v0.2.0..HEAD`)
+- Whether to push `main`
+- Whether to publish to PyPI now
+- Whether to create GitHub release now
+
+## Versioning requirement
+
+- Vtx's update check only supports strict numeric `MAJOR.MINOR.PATCH` versions such as `0.2.7` and `0.3.0`.
+- Do **not** cut releases with PyPI/PEP 440 prerelease or suffix forms like `0.3.0rc1`, `0.3.0b1`, `0.3.0.post1`, or `0.3.0.dev1` unless the update-check logic is updated first.
+- Even though PyPI commonly allows those formats, Vtx releases should continue following plain `X.Y.Z` so update detection stays correct.
+
+## Files to bump
+
+- `pyproject.toml` → `[project].version`
+- `src/vtx/version.py` → fallback `VERSION = "..."`
+- `uv.lock` → local package version block
+
+## Release workflow
+
+1. **Preflight**
+   - `git status --short --branch` must be clean (or confirm with user)
+   - `git tag --list` and `git log --oneline <prev_tag>..HEAD` to summarize changes
+
+2. **Update CHANGELOG.md**
+   - Replace the `## [Unreleased]` section's `- No changes yet.` with a new versioned heading: `## <version> - YYYY-MM-DD`
+   - Use `git log --oneline <prev_tag>..HEAD` to categorize changes into `### Added`, `### Changed`, `### Fixed` sections
+   - Credit external contributors with `- @username`
+   - Commit message: `docs: update changelog for <version>`
+
+3. **Version bump**
+   - Update version in all 3 files above
+
+4. **Quality gates**
+   - `uv run ruff format .`
+   - `uv run ruff check .`
+   - `uvx ty check .`
+   - `uv run python -m pytest`
+
+5. **Commit**
+   - Commit message: `build: bump version to <version>`
+
+6. **Tag**
+   - Annotated tag: `git tag -a v<version> -m "v<version> ..."`
+   - Include concise "changes since previous tag" bullets
+
+7. **Push**
+   - `git push vtx main`
+   - `git push vtx v<version>`
+
+8. **Build + verify artifacts**
+   - `rm -rf dist && uv build`
+   - `uv run python -m twine check dist/*`
+
+9. **Publish to PyPI**
+   - Prefer token file if present (example `~/.pypi-token`):
+   - `TWINE_USERNAME=__token__ TWINE_PASSWORD="$(< ~/.pypi-token)" uv run python -m twine upload dist/*`
+   - Verify:
+     - `https://pypi.org/project/vtx-coding-agent/<version>/`
+     - `https://pypi.org/pypi/vtx-coding-agent/json` reports latest version
+
+10. **Create GitHub release**
+   - If token exists at `~/.github-token`, call Releases API:
+   - `POST /repos/<owner>/<repo>/releases` with:
+     - `tag_name: v<version>`
+     - `target_commitish: main`
+     - `name: v<version>`
+     - `generate_release_notes: true`
+   - If 403 occurs, report missing token scopes/permissions (`contents:write` required)
+
+## Important notes
+
+- **Tagging and GitHub release are separate**:
+  - Tag = git ref in repository
+  - Release = GitHub object attached to a tag (notes/assets)
+- You can do either independently, but most projects do both together for user-facing releases.
+- If PyPI publish succeeds but GitHub release fails, do **not** retag/re-publish. Just fix auth and create the release for the existing tag.
+
+## Output checklist to report
+
+- Changelog updated for `<version>`
+- Version bumped in all files
+- Checks passed
+- Commit hash
+- Tag created and pushed
+- PyPI upload URL
+- GitHub release URL (or exact error + remediation)

vtx_coding_agent-0.1.1/.agents/skills/vtx-tmux-test/SKILL.md

@@ -0,0 +1,250 @@
+---
+name: vtx-tmux-test
+description: "E2E testing of vtx using tmux sessions; IMPORTANT: only trigger this skill when user asks for e2e testing of vtx"
+---
+
+# Vtx Tmux E2E Testing
+
+End-to-end testing of vtx using tmux sessions to programmatically control the TUI application.
+
+## Why Tmux?
+
+Vtx is a TUI (Textual-based) app. Running tests programmatically is hard. Tmux provides:
+
+- `tmux new-session` - isolate test environment
+- `tmux send-keys` - send keyboard input
+- `tmux capture-pane` - capture output
+- `tmux has-session` - check if vtx is running
+
+## Test Design Philosophy
+
+- **Deterministic**: Shell scripts create reproducible test environments
+- **Isolated config**: Tests run with `HOME=/tmp/vtx-e2e-home` so runtime settings do not mutate the real user config; auth JSON files are copied into the temp HOME when present so provider startup still works
+- **Separation of concerns**: Shell script runs steps and captures output; vtx/the reviewer evaluates results
+- **Output-based evaluation**: Test success/failure determined by reading output files, not shell script heuristics
+- **UI-focused**: Test triggers (`@`, `/`, runtime pickers, keybindings) by checking UI elements appear
+- **Filesystem verification**: Tool execution is verified through files under `/tmp/vtx-test-project`
+
+## Quick Start
+
+```bash
+# Run all e2e tests from the repo root
+bash .agents/skills/vtx-tmux-test/run-e2e-tests.sh
+
+# Optional: keep the temporary HOME for debugging
+KEEP_E2E_HOME=1 bash .agents/skills/vtx-tmux-test/run-e2e-tests.sh
+
+# Optional: override launch command/provider/model
+VTX_CMD='uv run vtx --model gpt-5.5' \
+  bash .agents/skills/vtx-tmux-test/run-e2e-tests.sh
+```
+
+After running, read `/tmp/vtx-test-*.txt` and evaluate the captured pane/config/filesystem outputs.
+
+## Test Scripts
+
+### Setup Script: `setup-test-project.sh`
+
+Creates a deterministic test project structure at `/tmp/vtx-test-project/`.
+
+```bash
+bash .agents/skills/vtx-tmux-test/setup-test-project.sh
+```
+
+### Main Test Script: `run-e2e-tests.sh`
+
+Runs comprehensive e2e tests including UI triggers, runtime controls, tab completion, and tool execution.
+
+```bash
+bash .agents/skills/vtx-tmux-test/run-e2e-tests.sh
+```
+
+## Test Categories
+
+### UI Trigger Tests (LLM-independent)
+
+- **/ commands**: Type `/`, verify slash command list appears with core and newer commands
+- **@ file search**: Type `@pyproject`, verify file picker appears with `pyproject.toml`
+- **/model command**: Type `/model`, verify model selector appears, then dismiss
+- **/new command**: Type `/new`, verify new conversation is started
+- **/resume command**: Type `/resume`, verify session list appears, then dismiss
+- **/session command**: Type `/session`, verify session info/statistics displayed
+
+### Runtime Mode Tests (LLM-independent)
+
+- **/permissions picker**: Shows `prompt` and `auto`, with current mode checked
+- **/permissions auto/prompt**: Info bar updates (`✓✓ auto` / `⏸ prompt`) and temp config persists `permissions.mode`
+- **Shift+Tab**: Cycles permission mode and persists it in temp config
+- **/thinking picker**: Shows `none`, `minimal`, `low`, `medium`, `high`, `xhigh`, with current level checked
+- **/thinking minimal**: Info bar model/thinking area updates to `minimal`
+- **Ctrl+Shift+T**: Cycles thinking level in the info bar
+- **/notifications picker**: Shows `on` and `off`, with current mode checked
+- **/notifications on/off**: Status says saved and temp config persists `notifications.enabled`
+- **Info bar row2 regression**: Permission mode remains row2-left while model/provider/thinking remains row2-right after runtime changes
+
+### Tab Path Completion Tests (LLM-independent)
+
+- **Unique match**: Type `pypr` + Tab, verify completes to `pyproject.toml`
+- **Multiple alternatives**: Type `src/vtx/ui/s` + Tab, verify floating list shows `selection_mode.py`, `session_ui.py`, `styles.py`
+- **Nested unique file**: Type `src/vtx/ui/widg` + Tab, verify completes to `src/vtx/ui/widgets.py`
+- **Select from list**: Type `src/vtx/ui/s` + Tab + Enter, verify first completion is applied to input
+
+### Tool Execution Tests (Filesystem verification)
+
+- **Write tool**: Creates `/tmp/vtx-test-project/test1.txt`, verified by file existence
+- **Edit tool**: Modifies `test1.txt`, verified by content changing from `hello` to `world`
+- **List files**: Shows directory contents in captured pane
+- **Calculation**: Computes `3+3`, verified in LLM output where practical
+
+## Configuration
+
+Edit or override environment variables for `run-e2e-tests.sh`:
+
+```bash
+WAIT_TIME=30                    # Time for LLM to complete all tool tasks
+COMMAND_WAIT_TIME=3             # Time for UI commands to settle
+SESSION_NAME="vtx-test"         # Tmux session name
+TEST_DIR="/tmp/vtx-test-project" # Test project directory for tool execution
+TEST_HOME="/tmp/vtx-e2e-home"    # Isolated HOME/config/session directory
+VTX_DIR="$PWD"                  # Vtx repo directory for tab completion tests
+VTX_CMD="uv run vtx --model gpt-5.5"
+KEEP_E2E_HOME=0                 # Set to 1 to preserve temp HOME after run
+```
+
+## Output Files
+
+The main script writes captured outputs to `/tmp/vtx-test-*.txt`:
+
+- `/tmp/vtx-test-1-commands.txt` - `/` slash command list
+- `/tmp/vtx-test-2-at-trigger.txt` - `@pyproject` file picker
+- `/tmp/vtx-test-3-model.txt` - `/model` selector
+- `/tmp/vtx-test-4-new.txt` - `/new` result
+- `/tmp/vtx-test-5-permissions-picker.txt` - `/permissions` picker
+- `/tmp/vtx-test-6-permissions-auto.txt` and `...-config.txt` - `/permissions auto`
+- `/tmp/vtx-test-7-permissions-prompt.txt` and `...-config.txt` - `/permissions prompt`
+- `/tmp/vtx-test-8-permissions-shift-tab.txt` and `...-config.txt` - Shift+Tab mode cycling
+- `/tmp/vtx-test-9-thinking-picker.txt` - `/thinking` picker
+- `/tmp/vtx-test-10-thinking-minimal.txt` - `/thinking minimal`
+- `/tmp/vtx-test-11-thinking-cycle.txt` - Ctrl+Shift+T thinking cycle
+- `/tmp/vtx-test-12-notifications-picker.txt` - `/notifications` picker
+- `/tmp/vtx-test-13-notifications-on.txt` and `...-config.txt` - `/notifications on`
+- `/tmp/vtx-test-14-notifications-off.txt` and `...-config.txt` - `/notifications off`
+- `/tmp/vtx-test-15-tab-unique.txt` - Tab completion unique match
+- `/tmp/vtx-test-16-tab-multiple.txt` - Tab completion alternatives
+- `/tmp/vtx-test-17-tab-nested-unique.txt` - Nested unique file completion
+- `/tmp/vtx-test-18-tab-select.txt` - Tab completion selection
+- `/tmp/vtx-test-19-tools.txt` - Tool execution turn
+- `/tmp/vtx-test-20-session.txt` - `/session` stats
+- `/tmp/vtx-test-21-resume.txt` - `/resume` session list
+- `/tmp/vtx-test-files.txt` - Test project file listing
+- `/tmp/vtx-test-test1-content.txt` - Final `test1.txt` content or `FILE_NOT_FOUND`
+- `/tmp/vtx-test-session-files.txt` - Session JSONL paths under temp HOME
+- `/tmp/vtx-test-final-config.txt` - Final temp config
+
+## Key Tmux Gotchas
+
+- **Use `Escape` not `Esc`**: tmux recognizes `Escape`. `Esc` sends literal characters.
+- **Always clear input between tests**: Use `Escape` to dismiss completions, then `C-u` to clear text.
+- **Completion selectors block input**: Selectors intercept Enter/Escape; dismiss them before the next test.
+- **Shift+Tab**: The script sends CSI Z via `Escape '[' 'Z'` rather than relying on a tmux key name.
+- **Ctrl+Shift+T**: The script sends CSI-u `Escape '[84;6u'` because `C-S-t` often collapses to Ctrl+T.
+
+## Test Evaluation (by Vtx/reviewer)
+
+After running the test script, evaluate results by reading the output files.
+
+### What to Check
+
+**UI Trigger Tests:**
+
+- `/` test: Slash command list includes `github`, `themes`, `permissions`, `thinking`, `notifications`, `init`, `compact`, `handoff`, `export`, `copy`, `login`, `logout`
+- `@` test: File picker appears and shows `pyproject.toml`
+- `/model` test: Model selector appears with model list/current markers
+- `/new` test: `Started new conversation` appears
+- `/resume` test: Session list appears with prior sessions
+- `/session` test: Session info/statistics displayed
+
+**Runtime Mode Tests:**
+
+- `/permissions` picker shows `prompt` and `auto`, current item checked
+- `/permissions auto` shows `✓✓ auto`, saved status, and config has `mode = "auto"`
+- `/permissions prompt` shows `⏸ prompt`, saved status, and config has `mode = "prompt"`
+- Shift+Tab toggles back to `auto` and config has `mode = "auto"`
+- `/thinking` picker shows `none`, `minimal`, `low`, `medium`, `high`, `xhigh`
+- `/thinking minimal` shows `Thinking level changed to minimal` and info bar row2-right includes `minimal`
+- Ctrl+Shift+T changes info bar thinking level from `minimal` to the next level
+- `/notifications` picker shows `on` and `off`, current item checked
+- `/notifications on/off` status says saved and config flips `enabled = true/false`
+- Permission mode remains in row2-left and model/provider/thinking remains row2-right
+
+**Tab Path Completion Tests:**
+
+- `pypr` + Tab shows `pyproject.toml`
+- `src/vtx/ui/s` + Tab shows `selection_mode.py`, `session_ui.py`, `styles.py`
+- `src/vtx/ui/widg` + Tab shows `src/vtx/ui/widgets.py`
+- `src/vtx/ui/s` + Tab + Enter applies a selected completion
+
+**Tool Execution Tests:**
+
+- `/tmp/vtx-test-project/test1.txt` exists
+- `/tmp/vtx-test-test1-content.txt` contains `world`
+- `/tmp/vtx-test-files.txt` lists `test1.txt`
+- `/tmp/vtx-test-19-tools.txt` shows relevant tool blocks/results
+
+### Tabular Report
+
+Provide a summary showing:
+
+- Test name
+- Status (PASS/FAIL)
+- Description/failure reason
+- Overall success rate
+
+### IMPORTANT: Always offer the view command
+
+After presenting the report, ALWAYS give the user this shell command so they can inspect raw captured outputs:
+
+```bash
+for f in /tmp/vtx-test-*.txt; do printf "\n\033[1;36m▶▶▶ %s\033[0m\n" "$f"; awk 'NF{found=1} found{lines[++n]=$0} END{while(n>0 && lines[n]=="") n--; for(i=1;i<=n;i++) print lines[i]}' "$f"; done
+```
+
+## Cleanup
+
+```bash
+# Test script auto-cleans tmux session and temp HOME unless KEEP_E2E_HOME=1.
+# Output files remain for evaluation (/tmp/vtx-test-*.txt).
+# Manual cleanup if needed:
+tmux kill-session -t vtx-test 2>/dev/null
+rm -rf /tmp/vtx-test-project /tmp/vtx-e2e-home
+rm -f /tmp/vtx-test-*.txt
+```
+
+## Tmux Commands Reference
+
+```bash
+# Session management
+tmux new-session -d -s <name> -c <dir> '<command>'
+tmux kill-session -t <name>
+tmux has-session -t <name>
+
+# Input — IMPORTANT: use full key names (Escape, Enter, not Esc)
+tmux send-keys -t <name> "text"
+tmux send-keys -t <name> Enter
+tmux send-keys -t <name> Escape
+tmux send-keys -t <name> Tab
+tmux send-keys -t <name> C-c
+tmux send-keys -t <name> C-u
+
+# Output
+tmux capture-pane -t <name> -p
+tmux capture-pane -t <name> -p > file.txt
+```
+
+## Tips
+
+- Tests are deterministic: project/config structure is recreated each run.
+- Runtime mode tests are LLM-independent and should be checked first.
+- Tab completion tests run from the vtx repo to use known paths.
+- Tool tests verify filesystem state; avoid relying solely on LLM prose.
+- Use `KEEP_E2E_HOME=1` to inspect temp config/session files after failures.
+- Run tool execution before `/resume` so there is a session with messages in the list.

vtx_coding_agent-0.1.1/.agents/skills/vtx-tmux-test/run-e2e-tests.sh

@@ -0,0 +1,312 @@
+#!/bin/bash
+
+# Comprehensive e2e test script for vtx
+# Tests UI triggers (@, / commands, runtime mode controls, Tab completion) and tool execution.
+# This script runs steps and captures output — evaluation is done by vtx reading the output files.
+
+set -u
+
+# Configuration
+WAIT_TIME=${WAIT_TIME:-30}                    # Time for LLM to complete tool tasks
+COMMAND_WAIT_TIME=${COMMAND_WAIT_TIME:-3}     # Time for UI commands to settle
+SESSION_NAME=${SESSION_NAME:-"vtx-test"}
+TEST_DIR=${TEST_DIR:-"/tmp/vtx-test-project"}
+TEST_HOME=${TEST_HOME:-"/tmp/vtx-e2e-home"}
+VTX_DIR=${VTX_DIR:-"$PWD"}                   # use caller's current working directory
+VTX_CMD=${VTX_CMD:-"uv run vtx --model gpt-5.5"}
+KEEP_E2E_HOME=${KEEP_E2E_HOME:-0}
+
+# Helper functions
+cleanup() {
+    tmux kill-session -t "$SESSION_NAME" 2>/dev/null || true
+    if [ "$KEEP_E2E_HOME" != "1" ]; then
+        rm -rf "$TEST_HOME"
+    fi
+}
+
+capture() {
+    tmux capture-pane -t "$SESSION_NAME" -p > "$1"
+}
+
+capture_config() {
+    if [ -f "$TEST_HOME/.vtx/config.yml" ]; then
+        cp "$TEST_HOME/.vtx/config.yml" "$1"
+    else
+        echo "CONFIG_NOT_FOUND" > "$1"
+    fi
+}
+
+# Dismiss any open completion/selector and clear the input line.
+# Uses Escape (NOT "Esc" which tmux would send as literal text).
+clear_input() {
+    tmux send-keys -t "$SESSION_NAME" Escape
+    sleep 0.5
+    tmux send-keys -t "$SESSION_NAME" Escape
+    sleep 0.5
+    tmux send-keys -t "$SESSION_NAME" C-u
+    sleep 0.5
+}
+
+run_command() {
+    local command="$1"
+    tmux send-keys -t "$SESSION_NAME" "$command"
+    sleep 1
+    tmux send-keys -t "$SESSION_NAME" Enter
+    sleep "$COMMAND_WAIT_TIME"
+}
+
+send_shift_tab() {
+    # tmux doesn't reliably expose a portable S-Tab key name, so send the terminal
+    # escape sequence directly: CSI Z.
+    tmux send-keys -t "$SESSION_NAME" Escape '[' 'Z'
+}
+
+send_ctrl_shift_t() {
+    # CSI-u encoding for Ctrl+Shift+T. This is more reliable in tmux than C-S-t,
+    # which collapses to Ctrl+T in many terminals.
+    tmux send-keys -t "$SESSION_NAME" Escape '[84;6u'
+}
+
+# Cleanup on exit
+trap cleanup EXIT
+
+# === Setup ===
+echo "Setting up isolated e2e environment..."
+cleanup
+rm -rf "$TEST_DIR" "$TEST_HOME"
+mkdir -p "$TEST_DIR" "$TEST_HOME/.vtx"
+if [ -f "$HOME/.vtx/openai_auth.json" ]; then
+    cp "$HOME/.vtx/openai_auth.json" "$TEST_HOME/.vtx/openai_auth.json"
+fi
+if [ -f "$HOME/.vtx/copilot_auth.json" ]; then
+    cp "$HOME/.vtx/copilot_auth.json" "$TEST_HOME/.vtx/copilot_auth.json"
+fi
+cd "$TEST_DIR" || exit 1
+printf '# Test Project\n' > README.md
+printf '{"name": "test"}\n' > config.json
+
+# Let vtx create its default config in the temp HOME. Runtime tests mutate only
+# this isolated file, never the real user config.
+
+# Clean up old test output files
+rm -f /tmp/vtx-test-*.txt
+
+# === Start vtx (from vtx repo for tab completion tests) ===
+echo "Starting vtx in tmux from vtx repo with HOME=$TEST_HOME..."
+tmux new-session -d -s "$SESSION_NAME" -c "$VTX_DIR" "HOME=$TEST_HOME OPENAI_API_KEY=\"${OPENAI_API_KEY:-}\" ZAI_API_KEY=\"${ZAI_API_KEY:-}\" ANTHROPIC_API_KEY=\"${ANTHROPIC_API_KEY:-}\" AZURE_AI_FOUNDRY_API_KEY=\"${AZURE_AI_FOUNDRY_API_KEY:-}\" AZURE_AI_FOUNDRY_BASE_URL=\"${AZURE_AI_FOUNDRY_BASE_URL:-}\" $VTX_CMD"
+sleep 5  # Give vtx time to start and render UI
+
+# =============================================================================
+# Test 1: / slash commands trigger
+# Verify: typing / shows the slash command list, including newer runtime commands
+# =============================================================================
+echo "Test 1: / slash commands trigger..."
+tmux send-keys -t "$SESSION_NAME" '/'
+sleep 2
+capture /tmp/vtx-test-1-commands.txt
+clear_input
+
+# =============================================================================
+# Test 2: @ file search trigger
+# Verify: typing @pyproject shows file picker with pyproject.toml (from vtx repo)
+# =============================================================================
+echo "Test 2: @ file search trigger..."
+tmux send-keys -t "$SESSION_NAME" '@pyproject'
+sleep 2
+capture /tmp/vtx-test-2-at-trigger.txt
+clear_input
+
+# =============================================================================
+# Test 3: /model command
+# Verify: /model shows model selector list, then dismiss without selecting
+# =============================================================================
+echo "Test 3: /model command..."
+tmux send-keys -t "$SESSION_NAME" '/model'
+sleep 2
+tmux send-keys -t "$SESSION_NAME" Enter
+sleep "$COMMAND_WAIT_TIME"
+capture /tmp/vtx-test-3-model.txt
+clear_input
+
+# =============================================================================
+# Test 4: /new command
+# Verify: /new starts a new conversation ("Started new conversation" message)
+# =============================================================================
+echo "Test 4: /new command..."
+run_command '/new'
+capture /tmp/vtx-test-4-new.txt
+
+# =============================================================================
+# P0 runtime mode controls and info bar
+# =============================================================================
+echo "Test 5: /permissions picker..."
+tmux send-keys -t "$SESSION_NAME" '/permissions'
+sleep 2
+tmux send-keys -t "$SESSION_NAME" Enter
+sleep "$COMMAND_WAIT_TIME"
+capture /tmp/vtx-test-5-permissions-picker.txt
+clear_input
+
+echo "Test 6: /permissions auto..."
+run_command '/permissions auto'
+capture /tmp/vtx-test-6-permissions-auto.txt
+capture_config /tmp/vtx-test-6-permissions-auto-config.txt
+
+echo "Test 7: /permissions prompt..."
+run_command '/permissions prompt'
+capture /tmp/vtx-test-7-permissions-prompt.txt
+capture_config /tmp/vtx-test-7-permissions-prompt-config.txt
+
+echo "Test 8: Shift+Tab permission cycling..."
+send_shift_tab
+sleep "$COMMAND_WAIT_TIME"
+capture /tmp/vtx-test-8-permissions-shift-tab.txt
+capture_config /tmp/vtx-test-8-permissions-shift-tab-config.txt
+
+echo "Test 9: /thinking picker..."
+tmux send-keys -t "$SESSION_NAME" '/thinking'
+sleep 2
+tmux send-keys -t "$SESSION_NAME" Enter
+sleep "$COMMAND_WAIT_TIME"
+capture /tmp/vtx-test-9-thinking-picker.txt
+clear_input
+
+echo "Test 10: /thinking minimal..."
+run_command '/thinking minimal'
+capture /tmp/vtx-test-10-thinking-minimal.txt
+
+echo "Test 11: Ctrl+Shift+T thinking cycling..."
+send_ctrl_shift_t
+sleep "$COMMAND_WAIT_TIME"
+capture /tmp/vtx-test-11-thinking-cycle.txt
+
+echo "Test 12: /notifications picker..."
+tmux send-keys -t "$SESSION_NAME" '/notifications'
+sleep 2
+tmux send-keys -t "$SESSION_NAME" Enter
+sleep "$COMMAND_WAIT_TIME"
+capture /tmp/vtx-test-12-notifications-picker.txt
+clear_input
+
+echo "Test 13: /notifications on..."
+run_command '/notifications on'
+capture /tmp/vtx-test-13-notifications-on.txt
+capture_config /tmp/vtx-test-13-notifications-on-config.txt
+
+echo "Test 14: /notifications off..."
+run_command '/notifications off'
+capture /tmp/vtx-test-14-notifications-off.txt
+capture_config /tmp/vtx-test-14-notifications-off-config.txt
+
+# =============================================================================
+# Test 15: Tab completion - unique match
+# Verify: typing "pypr" then Tab completes to "pyproject.toml"
+# =============================================================================
+echo "Test 15: Tab completion - unique match..."
+tmux send-keys -t "$SESSION_NAME" 'pypr'
+sleep 1
+tmux send-keys -t "$SESSION_NAME" Tab
+sleep 2
+capture /tmp/vtx-test-15-tab-unique.txt
+clear_input
+
+# =============================================================================
+# Test 16: Tab completion - multiple alternatives (floating list)
+# Verify: typing "src/vtx/ui/s" then Tab shows a list including:
+# selection_mode.py, session_ui.py, styles.py
+# =============================================================================
+echo "Test 16: Tab completion - multiple alternatives..."
+tmux send-keys -t "$SESSION_NAME" 'src/vtx/ui/s'
+sleep 1
+tmux send-keys -t "$SESSION_NAME" Tab
+sleep 2
+capture /tmp/vtx-test-16-tab-multiple.txt
+clear_input
+
+# =============================================================================
+# Test 17: Tab completion - nested unique file
+# Verify: typing "src/vtx/ui/widg" then Tab completes to "src/vtx/ui/widgets.py"
+# =============================================================================
+echo "Test 17: Tab completion - nested unique file..."
+tmux send-keys -t "$SESSION_NAME" 'src/vtx/ui/widg'
+sleep 1
+tmux send-keys -t "$SESSION_NAME" Tab
+sleep 2
+capture /tmp/vtx-test-17-tab-nested-unique.txt
+clear_input
+
+# =============================================================================
+# Test 18: Tab completion - select from list
+# Verify: typing "src/vtx/ui/s" Tab shows list, then Enter applies first completion
+# =============================================================================
+echo "Test 18: Tab completion - select from list..."
+tmux send-keys -t "$SESSION_NAME" 'src/vtx/ui/s'
+sleep 1
+tmux send-keys -t "$SESSION_NAME" Tab
+sleep 2
+tmux send-keys -t "$SESSION_NAME" Enter
+sleep 1
+capture /tmp/vtx-test-18-tab-select.txt
+clear_input
+
+# =============================================================================
+# Test 19: Tool execution (multiple tool calls)
+# Verify: creates test1.txt, edits it, lists files, calculates 3+3
+# Running this BEFORE /resume so there's a session with messages to resume.
+# Permission mode is auto from Shift+Tab above, so approval prompts should not block.
+# =============================================================================
+echo "Test 19: Tool execution..."
+run_command '/new'
+tmux send-keys -t "$SESSION_NAME" "Create $TEST_DIR/test1.txt containing \"hello\", then edit $TEST_DIR/test1.txt to change \"hello\" to \"world\", list files in $TEST_DIR, and calculate 3+3. Use parallel tool calls, be quick."
+sleep 1
+tmux send-keys -t "$SESSION_NAME" Enter
+sleep "$WAIT_TIME"
+capture /tmp/vtx-test-19-tools.txt
+
+# =============================================================================
+# Test 20: /session command
+# Verify: shows session info (messages, tokens, file path)
+# =============================================================================
+echo "Test 20: /session command..."
+run_command '/session'
+capture /tmp/vtx-test-20-session.txt
+
+# =============================================================================
+# Test 21: /resume command
+# Verify: shows list of sessions (at least one from tool execution above)
+# =============================================================================
+echo "Test 21: /resume command..."
+tmux send-keys -t "$SESSION_NAME" '/resume'
+sleep 2
+tmux send-keys -t "$SESSION_NAME" Enter
+sleep "$COMMAND_WAIT_TIME"
+capture /tmp/vtx-test-21-resume.txt
+clear_input
+
+# =============================================================================
+# Capture file system/session/config state for verification
+# =============================================================================
+echo "Capturing file system and persisted state..."
+ls -la "$TEST_DIR" > /tmp/vtx-test-files.txt 2>/dev/null
+find "$TEST_HOME/.vtx/sessions" -type f -name '*.jsonl' -print > /tmp/vtx-test-session-files.txt 2>/dev/null || true
+capture_config /tmp/vtx-test-final-config.txt
+
+# Retry a few times in case the LLM is still finishing file writes
+for _ in 1 2 3; do
+    if [ -f "$TEST_DIR/test1.txt" ]; then
+        cat "$TEST_DIR/test1.txt" > /tmp/vtx-test-test1-content.txt
+        break
+    fi
+    sleep 3
+done
+# Final check
+if [ ! -f "$TEST_DIR/test1.txt" ]; then
+    echo "FILE_NOT_FOUND" > /tmp/vtx-test-test1-content.txt
+    ls -la "$TEST_DIR" > /tmp/vtx-test-files.txt 2>/dev/null
+fi
+
+printf '\n%s\n' "${SEP:-==============================}"
+echo "All tests complete"
+echo "Output files saved to /tmp/vtx-test-*.txt"
+echo "Temp HOME: $TEST_HOME (KEEP_E2E_HOME=1 to keep after run)"
+printf '%s\n' "${SEP:-==============================}"

vtx_coding_agent-0.1.1/.agents/skills/vtx-tmux-test/setup-test-project.sh

@@ -0,0 +1,27 @@
+#!/bin/bash
+
+# Setup script for vtx e2e testing
+# Creates a deterministic test project structure at /tmp/vtx-test-project
+
+TEST_DIR="/tmp/vtx-test-project"
+
+# Cleanup and recreate test project
+rm -rf $TEST_DIR
+mkdir -p $TEST_DIR
+cd $TEST_DIR
+
+# Create deterministic test files
+echo "# Test Project" > README.md
+echo '{"name": "test", "version": "1.0.0"}' > config.json
+echo "Todo:
+- Read config.json
+- Calculate sum" > notes.txt
+echo 'def add(a, b): return a + b
+def multiply(a, b): return a * b
+result = multiply(5, 3)
+print(f"5 * 3 = {result}")' > utils.py
+echo 'DATA = ["item1", "item2", "item3"]
+for item in DATA:
+    print(item)' > data.py
+
+echo "✓ Test project created at $TEST_DIR"