home / skills / antinomyhq / forge / debug-cli

debug-cli skill

/.forge/skills/debug-cli

This skill provides a structured workflow to debug and verify changes to the forge CLI, speeding fixes without committing.

npx playbooks add skill antinomyhq/forge --skill debug-cli

Review the files below or copy the command above to add this skill to your agents.

Files (3)
SKILL.md
6.4 KB
---
name: debug-cli
description: Use when users need to debug, modify, or extend the code-forge application's CLI commands, argument parsing, or CLI behavior. This includes adding new commands, fixing CLI bugs, updating command options, or troubleshooting CLI-related issues.
---

# CLI Debug Skill

This skill provides a systematic workflow for debugging and verifying changes to the forge CLI application.

## Core Principles

1. **Always get latest docs first**: Run `--help` to see current commands and options
2. **Use `-p` for testing**: Test forge by giving it tasks with the `-p` flag
3. **Never commit**: This is for debugging only - don't commit changes
4. **Clone conversations**: When debugging conversation bugs, clone the source conversation before reproducing

## Workflow

### 1. Build the Application

Always build in debug mode after making changes:

```bash
cargo build
```

**Never** use `cargo build --release` for debugging - it's significantly slower and unnecessary for verification.

### 2. Get Latest Documentation

**Always** start by checking the latest help to understand current commands and options:

```bash
# Main help - do this first
./target/debug/forge --help

# Command-specific help
./target/debug/forge [COMMAND] --help

# Subcommand help
./target/debug/forge [COMMAND] [SUBCOMMAND] --help
```

### 3. Test with `-p` Flag

Use the `-p` flag to give forge a task to complete without interactive mode:

```bash
# Test with a simple prompt
./target/debug/forge -p "create a hello world rust program"

# Test with specific functionality
./target/debug/forge -p "read the README.md file and summarize it"

# Test with complex tasks
./target/debug/forge -p "analyze the code structure and suggest improvements"
```

### 4. Debug with Conversation Dumps

When debugging prompts or conversation issues, use `conversation dump` to export conversations. The command automatically creates a timestamped file:

```bash
# Dump conversation as JSON (creates: YYYY-MM-DD_HH-MM-SS-dump.json)
./target/debug/forge conversation dump <conversation-id>

# Export as HTML for human-readable format (creates: YYYY-MM-DD_HH-MM-SS-dump.html)
./target/debug/forge conversation dump --html <conversation-id>

# Use dumped JSON to reproduce issues
./target/debug/forge --conversation 2025-11-23_12-28-52-dump.json
```

### 5. Clone Before Reproducing Bugs

**Critical**: When a user provides a conversation with a bug, always clone it first:

```bash
# Clone the conversation
./target/debug/forge conversation clone <source-conversation-id>

# This creates a new conversation ID - use that for testing
./target/debug/forge --conversation-id <new-cloned-id>

# Keep cloning the source until the fix is verified
# Never modify the original conversation
```

**Why clone?**

- Preserves original bug evidence
- Allows multiple reproduction attempts
- Enables A/B testing of fixes
- Keeps source conversation clean

## Common Testing Patterns

### Test New Features

```bash
# Build and test new command
cargo build
./target/debug/forge --help  # Verify new command appears
./target/debug/forge new-command --help  # Check command docs
./target/debug/forge -p "test the new feature"
```

### Reproduce Reported Bugs

```bash
# 1. Dump the conversation (creates timestamped JSON file)
./target/debug/forge conversation dump <bug-conversation-id>

# 2. Clone it for testing (preserves original)
./target/debug/forge conversation clone <bug-conversation-id>

# 3. Reproduce with the cloned conversation
./target/debug/forge --conversation-id <cloned-id> -p "reproduce the issue"

# 4. After fix, verify with new clone
./target/debug/forge conversation clone <bug-conversation-id>
./target/debug/forge --conversation-id <new-clone-id> -p "verify fix"
```

### Test Edge Cases

```bash
# Test with missing arguments
./target/debug/forge command

# Test with invalid input
./target/debug/forge -p "invalid task with special chars: <>|&"

# Test with boundary values
./target/debug/forge -p "create a file with a very long name..."
```

### Debug Prompt Optimization

```bash
# 1. Dump conversation to analyze prompts (creates timestamped JSON)
./target/debug/forge conversation dump <id>

# 2. Review the conversation structure
cat 2025-11-23_12-28-52-dump.json | jq '.messages[] | {role, content}'

# 3. Export as HTML for easier reading
./target/debug/forge conversation dump --html <id>

# 4. Test modified prompts
./target/debug/forge -p "your optimized prompt here"
```

## Integration with Development Workflow

### After Code Changes

1. **Build**: `cargo build`
2. **Docs**: `./target/debug/forge --help` (verify documentation)
3. **Test**: `./target/debug/forge -p "relevant task"`
4. **Verify**: Check output matches expectations

### Debugging a Bug Report

1. **Clone**: `./target/debug/forge conversation clone <source-id>`
2. **Build**: `cargo build` (with potential fixes)
3. **Test**: `./target/debug/forge --conversation-id <cloned-id> -p "reproduce"`
4. **Iterate**: Repeat until verified
5. **Never commit** during debugging - only after full verification

## Quick Reference

```bash
# Standard debug workflow
cargo build
./target/debug/forge --help  # Always check docs first
./target/debug/forge -p "your test task"

# Dump conversation (creates timestamped file)
./target/debug/forge conversation dump <id>
# Output: 2025-11-23_12-28-52-dump.json

# Export as HTML for review
./target/debug/forge conversation dump --html <id>
# Output: 2025-11-23_12-28-52-dump.html

# Use dumped conversation
./target/debug/forge --conversation 2025-11-23_12-28-52-dump.json

# Clone and test bug
./target/debug/forge conversation clone <source-id>
./target/debug/forge --conversation-id <cloned-id> -p "reproduce bug"

# Debug prompts with jq (use actual filename)
cat 2025-11-23_12-28-52-dump.json | jq '.messages[] | {role, content}'

# Test with verbose output
./target/debug/forge --verbose -p "test task"
```

## Tips

- **Always `--help` first**: Get latest docs before testing
- **Use `-p` for testing**: Don't test interactively, use prompts
- **Clone conversations**: Never modify original bug conversations
- **Never commit**: This is for debugging only
- **Dump creates files**: `dump` automatically creates timestamped files (no `>` needed)
- **HTML exports**: Use `--html` flag for human-readable conversation views
- **Use relative paths**: Binary is at `./target/debug/forge` from project root
- **Check exit codes**: Use `echo $?` to verify exit codes
- **Watch for warnings**: Build warnings often indicate issues

Overview

This skill provides a compact, practical workflow for debugging, extending, and verifying the forge CLI application. It focuses on safe iterative testing: build in debug mode, inspect help output, reproduce issues with conversation dumps, and clone conversations before modifying anything. The guidance is tuned for a Rust-based CLI development loop and emphasizes reproducibility and non-destructive testing.

How this skill works

The skill guides you to build the binary in debug mode, use --help to read current commands, and run tasks non-interactively with the -p flag to verify behavior. For conversation-related bugs it exports timestamped dumps (JSON or HTML), lets you clone the source conversation for safe reproduction, and shows patterns for reproducing and validating fixes without touching original data. It also covers edge-case tests and prompt optimization checks.

When to use it

  • Adding or updating CLI commands, flags, or argument parsing
  • Reproducing and diagnosing CLI behavior or exit-code issues
  • Debugging conversation-handling bugs or LLM prompt regressions
  • Testing new features quickly without committing changes
  • Verifying fixes with safe, cloned conversation data

Best practices

  • Always build with cargo build (debug mode) when testing changes
  • Run ./target/debug/forge --help first to capture current docs
  • Use -p for non-interactive tests; avoid interactive runs for repeatability
  • Dump and clone conversations before reproducing bugs to preserve evidence
  • Never commit debug-only changes; iterate until the fix is verified

Example use cases

  • Add a new subcommand: build, verify it appears in --help, then test with -p
  • Reproduce a reported LLM prompt bug by dumping the conversation and running with --conversation <dump.json>
  • Fix argument parsing: run malformed inputs and boundary values to confirm errors and exit codes
  • Optimize prompts by exporting conversation JSON, inspecting messages with jq, and testing modified prompts with -p
  • Verify a fix by cloning the original conversation and rerunning the reproduction steps

FAQ

Why build in debug mode instead of release?

Debug builds are faster to iterate on and easier to debug; release builds optimize and can obscure debugging details.

How do I preserve the original conversation when testing?

Use conversation dump to export the source, then conversation clone to create a working copy and run tests against the clone.