kon-coding-agent 0.1.0__tar.gz

kon_coding_agent-0.1.0/.gitignore

@@ -0,0 +1,19 @@
+# Python-generated files
+__pycache__/
+*.py[oc]
+build/
+dist/
+wheels/
+*.egg-info
+
+# Virtual environments
+.venv
+
+# macOS
+.DS_Store
+
+# Exported kon sessions
+kon-session-*
+
+# Local editor settings
+.vscode/

kon_coding_agent-0.1.0/.kon/skills/kon-tmux-test/SKILL.md

@@ -0,0 +1,208 @@
+---
+name: kon-tmux-test
+description: E2E testing of kon using tmux sessions; IMPORTANT: only trigger this skill when user asks for e2e testing of kon
+---
+
+# Kon Tmux E2E Testing
+
+End-to-end testing of kon using tmux sessions to programmatically control the TUI application.
+
+## Why Tmux?
+
+Kon 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 kon is running
+
+## Test Design Philosophy
+
+- **Deterministic**: Shell scripts create reproducible test environments
+- **Separation of concerns**: Shell script runs steps and captures output; kon (AI) evaluates results
+- **Output-based evaluation**: Test success/failure determined by AI reading output files, not shell script heuristics
+- **UI-focused**: Test triggers (@, / commands) by checking UI elements appear
+- **Single conversation**: Encourage all tool calls in one query
+
+## Quick Start
+
+```bash
+# Run all e2e tests
+bash ~/.kon/skills/kon-tmux-test/run-e2e-tests.sh
+
+# After running, kon (the AI) reads the output files and evaluates results
+# Output files: /tmp/kon-test-*.txt
+```
+
+## Test Scripts
+
+### Setup Script: `setup-test-project.sh`
+
+Creates the same deterministic test project structure each time at `/tmp/kon-test-project/`:
+
+```bash
+bash ~/.kon/skills/kon-tmux-test/setup-test-project.sh
+```
+
+Test project structure:
+```
+/tmp/kon-test-project/
+├── README.md          # Project documentation
+├── config.json       # Config file with basic data
+├── notes.txt        # Simple text notes
+├── utils.py         # Python utility file
+└── data.py          # Python data file
+```
+
+### Main Test Script: `run-e2e-tests.sh`
+
+Runs comprehensive e2e tests including UI triggers and tool execution:
+
+```bash
+bash ~/.kon/skills/kon-tmux-test/run-e2e-tests.sh
+```
+
+## Test Categories
+
+### UI Trigger Tests (LLM-independent)
+- **/ commands**: Type `/`, verify slash command list appears with all 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
+
+### Tab Path Completion Tests (LLM-independent)
+- **Unique match**: Type `sr` + Tab, verify completes to `src/` directly
+- **Multiple alternatives**: Type `s` + Tab, verify floating list shows `scripts/`, `src/`
+- **Home directory**: Type `~/De` + Tab, verify floating list shows `Desktop/`, `Developer/`, etc.
+- **Select from list**: Type `s` + Tab + Enter, verify `scripts/` is applied to input
+
+### Tool Execution Tests (File system verification)
+- **Write tool**: Creates files, verified by file existence
+- **Edit tool**: Modifies files, verified by file content
+- **List files**: Shows directory contents, verified by output
+- **Calculation**: Computes results, verified by LLM output (acceptable for non-tool operations)
+
+## Running Tests
+
+```bash
+# Run all tests - script executes steps and captures output
+bash ~/.kon/skills/kon-tmux-test/run-e2e-tests.sh
+
+# After running, the AI (kon) reads output files and provides evaluation:
+# - /tmp/kon-test-1-commands.txt       (/ slash commands list)
+# - /tmp/kon-test-2-at-trigger.txt     (@ file picker test)
+# - /tmp/kon-test-3-model.txt          (/model selector test)
+# - /tmp/kon-test-4-new.txt            (/new conversation test)
+# - /tmp/kon-test-5-tab-unique.txt     (Tab completion unique match)
+# - /tmp/kon-test-6-tab-multiple.txt   (Tab completion floating list)
+# - /tmp/kon-test-7-tab-home.txt       (Tab completion home directory)
+# - /tmp/kon-test-8-tab-select.txt     (Tab completion select from list)
+# - /tmp/kon-test-9-tools.txt          (tool execution test)
+# - /tmp/kon-test-10-session.txt       (/session stats test)
+# - /tmp/kon-test-11-resume.txt        (/resume session list test)
+# - /tmp/kon-test-files.txt            (file system state)
+# - /tmp/kon-test-test1-content.txt    (test1.txt content)
+```
+
+## Configuration
+
+Edit `run-e2e-tests.sh` to adjust:
+
+```bash
+WAIT_TIME=30          # Time for LLM to complete all tasks (adjust based on model speed)
+COMMAND_WAIT_TIME=3   # Time for UI commands to settle
+SESSION_NAME="kon-test"  # Tmux session name
+TEST_DIR="/tmp/kon-test-project"  # Test project directory for tool execution
+KON_DIR="$HOME/Developer/personal/kon"  # Kon repo directory (for tab completion tests)
+```
+
+## Key Tmux Gotchas
+
+- **Use `Escape` not `Esc`**: tmux recognizes `Escape` as the escape key. `Esc` is NOT a valid key name and sends literal characters 'E', 's', 'c' as text.
+- **Always clear input between tests**: Use `Escape` to dismiss completions, then `C-u` to clear text. Without this, text from one test bleeds into the next.
+- **Completion selectors block input**: The model selector and session list intercept Enter/Escape. Always dismiss them with `Escape` before the next test.
+
+## Test Evaluation (by Kon)
+
+After running the test script, kon evaluates the results by reading the output files:
+
+### What to Check
+
+**UI Trigger Tests (by reading output files):**
+- `/` test: Slash command list appears showing commands (help, model, new, etc.)
+- `@` test: File picker appears and shows files (pyproject.toml)
+- `/model` test: Model selector appears with model list
+- `/new` test: "Started new conversation" message appears
+- `/resume` test: Session list appears with prior sessions
+- `/session` test: Session info/statistics displayed (messages, tokens)
+
+**Tab Path Completion Tests (by reading output files):**
+- `sr` + Tab test: Input shows `src/` (unique completion applied)
+- `s` + Tab test: Floating list visible with `scripts/` and `src/` options
+- `~/De` + Tab test: Floating list visible with `Desktop/`, `Developer/`, `Documents/`, `Downloads/`
+- `s` + Tab + Enter test: Input shows `scripts/` (selection applied)
+
+**Tool Execution Tests (by reading output files + file system):**
+- `test1.txt` created in /tmp/kon-test-project/
+- `test1.txt` contains "world" (not "hello")
+- Directory listing shows files
+
+### Tabular Report
+
+Kon provides a tabular summary after reading all output files, showing:
+- Test name
+- Status (PASS/FAIL)
+- Description
+- Overall success rate
+
+### IMPORTANT: Always offer the view command
+
+After presenting the tabular report, ALWAYS give the user this shell command so they can inspect the raw captured outputs in their terminal:
+
+```bash
+for f in /tmp/kon-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 on exit via trap
+# Output files remain for kon to evaluate (/tmp/kon-test-*.txt)
+# Manual cleanup if needed:
+tmux kill-session -t kon-test 2>/dev/null
+rm -rf /tmp/kon-test-project
+rm -f /tmp/kon-test-*.txt
+```
+
+## Tmux Commands Reference
+
+```bash
+# Session management
+tmux new-session -d -s <name> -c <dir> '<command>'  # Create detached session
+tmux kill-session -t <name>                        # Kill session
+tmux has-session -t <name>                         # Check if session exists
+
+# Input — IMPORTANT: use full key names (Escape, Enter, not Esc)
+tmux send-keys -t <name> "text"    # Send text
+tmux send-keys -t <name> Enter     # Send Enter key
+tmux send-keys -t <name> Escape    # Send Escape key (NOT "Esc"!)
+tmux send-keys -t <name> Tab       # Send Tab key
+tmux send-keys -t <name> C-c       # Send Ctrl+C
+tmux send-keys -t <name> C-u       # Send Ctrl+U (clear line)
+
+# Output
+tmux capture-pane -t <name> -p     # Capture pane to stdout
+tmux capture-pane -t <name> -p > file.txt  # Save to file
+```
+
+## Tips
+
+- Tests are deterministic: same project structure recreated each run
+- UI tests don't depend on LLM: verify TUI elements appear
+- Tab completion tests run from kon repo to test on known directory structure
+- Tool tests verify file system: check actual files, not LLM output
+- Use `trap cleanup EXIT` to ensure tmux session is always cleaned up
+- Adjust `WAIT_TIME` based on model speed for tool execution tests
+- Output files saved to `/tmp/kon-test-*.txt` for debugging
+- Run tool execution before /resume so there's a session with messages in the list

kon_coding_agent-0.1.0/.kon/skills/kon-tmux-test/run-e2e-tests.sh

@@ -0,0 +1,222 @@
+#!/bin/bash
+
+# Comprehensive e2e test script for kon
+# Tests UI triggers (@, / commands, Tab completion) and tool execution
+# This script runs steps and captures output — evaluation is done by kon reading the output files
+
+# Configuration
+WAIT_TIME=30          # Time for LLM to complete tool tasks
+COMMAND_WAIT_TIME=3   # Time for UI commands to settle
+SESSION_NAME="kon-test"
+TEST_DIR="/tmp/kon-test-project"
+KON_DIR="$PWD"  # use caller's current working directory for tab completion tests
+
+# Helper functions
+cleanup() {
+    tmux kill-session -t "$SESSION_NAME" 2>/dev/null || true
+}
+
+capture() {
+    tmux capture-pane -t "$SESSION_NAME" -p > "$1"
+}
+
+# Dismiss any open completion/selector and clear the input line.
+# Uses Escape (NOT "Esc" which tmux would send as literal text).
+clear_input() {
+    tmux send-keys -t "$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
+}
+
+# Cleanup on exit
+trap cleanup EXIT
+
+# === Setup ===
+echo "Setting up test project..."
+rm -rf "$TEST_DIR"
+mkdir -p "$TEST_DIR"
+cd "$TEST_DIR" || exit 1
+echo "# Test Project" > README.md
+echo '{"name": "test"}' > config.json
+
+# Clean up old test output files
+rm -f /tmp/kon-test-*.txt
+
+# Clean up old sessions from previous runs, keep only the last 5
+SESSION_DIR="$HOME/.kon/sessions/private-tmp-kon-test-project"
+if [ -d "$SESSION_DIR" ]; then
+    ls -t "$SESSION_DIR"/*.jsonl 2>/dev/null | tail -n +6 | xargs rm -f 2>/dev/null
+fi
+
+# === Start kon (from kon repo for tab completion tests) ===
+echo "Starting kon in tmux from kon repo..."
+cleanup
+tmux new-session -d -s "$SESSION_NAME" -c "$KON_DIR" 'uv run kon'
+sleep 5  # Give kon time to start and render UI
+
+# =============================================================================
+# Test 1: / slash commands trigger
+# Verify: typing / shows the slash command list
+# =============================================================================
+echo "Test 1: / slash commands trigger..."
+tmux send-keys -t "$SESSION_NAME" '/'
+sleep 2
+capture /tmp/kon-test-1-commands.txt
+clear_input
+
+# =============================================================================
+# Test 2: @ file search trigger
+# Verify: typing @config shows file picker with config.json (from kon repo)
+# =============================================================================
+echo "Test 2: @ file search trigger..."
+tmux send-keys -t "$SESSION_NAME" '@pyproject'
+sleep 2
+capture /tmp/kon-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 "$SESSION_NAME" '/model'
+sleep 2
+# Enter applies the slash command autocomplete -> opens model selector
+tmux send-keys -t "$SESSION_NAME" Enter
+sleep "$COMMAND_WAIT_TIME"
+capture /tmp/kon-test-3-model.txt
+# Dismiss model selector without selecting (Escape hides the completion list)
+clear_input
+
+# =============================================================================
+# Test 4: /new command
+# Verify: /new starts a new conversation ("Started new conversation" message)
+# =============================================================================
+echo "Test 4: /new command..."
+tmux send-keys -t "$SESSION_NAME" '/new'
+sleep 2
+tmux send-keys -t "$SESSION_NAME" Enter
+sleep "$COMMAND_WAIT_TIME"
+capture /tmp/kon-test-4-new.txt
+
+# =============================================================================
+# Test 5: Tab completion - unique match
+# Verify: typing "sr" then Tab completes to "src/"
+# =============================================================================
+echo "Test 5: Tab completion - unique match..."
+tmux send-keys -t "$SESSION_NAME" 'sr'
+sleep 1
+tmux send-keys -t "$SESSION_NAME" Tab
+sleep 2
+capture /tmp/kon-test-5-tab-unique.txt
+clear_input
+
+# =============================================================================
+# Test 6: Tab completion - multiple alternatives (floating list)
+# Verify: typing "s" then Tab shows floating list with scripts/, src/
+# =============================================================================
+echo "Test 6: Tab completion - multiple alternatives..."
+tmux send-keys -t "$SESSION_NAME" 's'
+sleep 1
+tmux send-keys -t "$SESSION_NAME" Tab
+sleep 2
+capture /tmp/kon-test-6-tab-multiple.txt
+clear_input
+
+# =============================================================================
+# Test 7: Tab completion - home directory
+# Verify: typing "~/De" then Tab shows floating list with Desktop/, Developer/, etc.
+# =============================================================================
+echo "Test 7: Tab completion - home directory..."
+tmux send-keys -t "$SESSION_NAME" '~/De'
+sleep 1
+tmux send-keys -t "$SESSION_NAME" Tab
+sleep 2
+capture /tmp/kon-test-7-tab-home.txt
+clear_input
+
+# =============================================================================
+# Test 8: Tab completion - select from list
+# Verify: typing "s" Tab shows list, then select with Enter applies completion
+# =============================================================================
+echo "Test 8: Tab completion - select from list..."
+tmux send-keys -t "$SESSION_NAME" 's'
+sleep 1
+tmux send-keys -t "$SESSION_NAME" Tab
+sleep 2
+# Select first item (scripts/) with Enter
+tmux send-keys -t "$SESSION_NAME" Enter
+sleep 1
+capture /tmp/kon-test-8-tab-select.txt
+clear_input
+
+# =============================================================================
+# Test 9: Tool execution (multiple tool calls)
+# Verify: creates test1.txt, edits it, lists files, calculates 3+3
+# Running this BEFORE /resume so there's a session with messages to resume
+# Switch to test project dir first
+# =============================================================================
+echo "Test 9: Tool execution..."
+# Change to test directory for tool execution
+tmux send-keys -t "$SESSION_NAME" "/new"
+sleep 1
+tmux send-keys -t "$SESSION_NAME" Enter
+sleep 2
+tmux send-keys -t "$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 "$WAIT_TIME"
+capture /tmp/kon-test-9-tools.txt
+
+# =============================================================================
+# Test 10: /session command
+# Verify: shows session info (messages, tokens, file path)
+# =============================================================================
+echo "Test 10: /session command..."
+tmux send-keys -t "$SESSION_NAME" '/session'
+sleep 2
+tmux send-keys -t "$SESSION_NAME" Enter
+sleep "$COMMAND_WAIT_TIME"
+capture /tmp/kon-test-10-session.txt
+
+# =============================================================================
+# Test 11: /resume command
+# Verify: shows list of sessions (at least one from tool execution above)
+# =============================================================================
+echo "Test 11: /resume command..."
+tmux send-keys -t "$SESSION_NAME" '/resume'
+sleep 2
+tmux send-keys -t "$SESSION_NAME" Enter
+sleep "$COMMAND_WAIT_TIME"
+capture /tmp/kon-test-11-resume.txt
+# Dismiss session list without selecting
+clear_input
+
+# =============================================================================
+# Capture file system state for tool execution verification
+# =============================================================================
+echo "Capturing file system state..."
+ls -la "$TEST_DIR" > /tmp/kon-test-files.txt 2>/dev/null
+
+# Retry a few times in case the LLM is still finishing file writes
+for i in 1 2 3; do
+    if [ -f "$TEST_DIR/test1.txt" ]; then
+        cat "$TEST_DIR/test1.txt" > /tmp/kon-test-test1-content.txt
+        break
+    fi
+    sleep 3
+done
+# Final check
+if [ ! -f "$TEST_DIR/test1.txt" ]; then
+    echo "FILE_NOT_FOUND" > /tmp/kon-test-test1-content.txt
+    ls -la "$TEST_DIR" > /tmp/kon-test-files.txt 2>/dev/null
+fi
+
+echo ""
+echo "$SEP"
+echo "All tests complete"
+echo "Output files saved to /tmp/kon-test-*.txt"
+echo "$SEP"

kon_coding_agent-0.1.0/.kon/skills/kon-tmux-test/setup-test-project.sh

@@ -0,0 +1,27 @@
+#!/bin/bash
+
+# Setup script for kon e2e testing
+# Creates a deterministic test project structure at /tmp/kon-test-project
+
+TEST_DIR="/tmp/kon-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"

kon_coding_agent-0.1.0/.python-version

@@ -0,0 +1 @@
+3.13

kon_coding_agent-0.1.0/AGENTS.md

@@ -0,0 +1,20 @@
+# Agent Guidelines
+
+## Code Style
+
+- Don't add trivial docstrings. Only add docstrings when explaining complex functionality.
+- This project uses `uv`. Run `uv run ruff format .` after editing or creating any files.
+
+## Testing
+
+- Use `uv run pytest` for testing in general; after edits/writes
+- If the user asks for e2e tests then run the kon-tmux e2e test if available
+
+## Committing code
+
+- If the user tells you to commit code, look at all the changes and create multile commits if needed bsaed on logical groupings
+
+## Pushing
+
+- If the user asks you to push code, run these first before doing so: `uv run ruff format .`, `uv run ruff check .`, `uv run pyright .` and `uv run pytest` in parallel (same tool call)
+- Only if these all pass without issues should you push otherwise report the warnings/errors back to user and ask for next steps

kon_coding_agent-0.1.0/LICENSE

@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2026 kuutsav
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.