Current Configuration (Auto-Captured)

Prompts Verbosity: !grep PROMPTS_VERBOSITY .claude/.ref-config 2>/dev/null | cut -d= -f2 || echo "major (default)"
Sources Verbosity: !grep SOURCES_VERBOSITY .claude/.ref-config 2>/dev/null | cut -d= -f2 || echo "all (default)"
Export Path: !grep EXPORT_PATH .claude/.ref-config 2>/dev/null | cut -d= -f2 || echo "exports/ (default)"
Config File: ![ -f .claude/.ref-config ] && echo "✓ Exists" || echo "✗ Not found"

Quick Example

/track:config prompts=all sources=all
# ✓ Configuration updated
# PROMPTS_VERBOSITY: major → all (all requests tracked)
# SOURCES_VERBOSITY: all (unchanged)

/track:config
# Interactive mode - uses AskUserQuestion

config - View or Update Tracking Configuration

View or update tracking verbosity and export configuration.

Usage

Interactive mode (uses AskUserQuestion):

/track:config

Direct mode (update immediately):

/track:config prompts=all
/track:config sources=off
/track:config prompts=major sources=all
/track:config export_path=paper/references/
/track:config prompts=all export_path=exports/

What it does

1. With no arguments - Interactive configuration:

   • Uses AskUserQuestion to present configuration options
   • Shows current settings
   • Allows selection of new values
   • Updates ./.claude/.ref-config
   • Shows new configuration

2. With arguments - Direct update:

   • Parses key=value pairs
   • Updates ./.claude/.ref-config
   • Validates setting values
   • Shows new configuration

Configuration options

PROMPTS_VERBOSITY

Controls what prompts are tracked to claude_usage/prompts.md:

• major (default) - Only significant multi-step academic/development work

  • Example: "Implement authentication system", "Debug complex algorithm"
  • Heuristic: Response >100 words or multiple tool uses
  • Best for: Academic research, project documentation

• all - Track every user request

  • Example: "What is X?", "Explain Y", "Fix typo in file.txt"
  • Best for: Complete session logs, detailed auditing

• minimal - Only explicitly user-requested tracking

  • Only tracks when user says "track this" or similar
  • Best for: Selective manual curation

• off - Disable prompt tracking completely

  • No entries added to claude_usage/prompts.md
  • Best for: Source-only tracking

SOURCES_VERBOSITY

Controls what sources are tracked to claude_usage/sources.md:

• all (default) - Track all WebSearch/WebFetch/Read/Grep operations

  • Every search logged automatically
  • Documentation reads included
  • Best for: Complete research audit trail

• off - Disable source tracking completely

  • No entries added to claude_usage/sources.md
  • Best for: Prompt-only tracking

EXPORT_PATH (new in v2.0)

Default directory for /track:export command output:

• exports/ (default) - Standard export directory
• Can be absolute or relative path
• Used when no output path specified in export command
• Examples: exports/, paper/references/, /tmp/tracking/

Example output

Viewing config (interactive mode):

Uses AskUserQuestion with three questions:

1. Prompts verbosity: major (default) | all | minimal | off
2. Sources verbosity: all (default) | off
3. Export path: exports/ (default) | paper/ | custom path

Updating config (direct mode):

✓ Configuration updated

PROMPTS_VERBOSITY: major → all
SOURCES_VERBOSITY: all (unchanged)
EXPORT_PATH: exports/ (unchanged)

New behavior:
- Every user request will be tracked to claude_usage/prompts.md
- All searches continue to be tracked to claude_usage/sources.md
- Exports default to exports/ directory

Prerequisites

Run /track:init first to create .claude/.ref-config.

Notes

• Changes take effect immediately
• Existing tracked data is not affected
• Works regardless of auto-tracking status (on/off)
• Invalid values are rejected with error message
• Interactive mode recommended for first-time configuration

v2.0 Changes

New in v2.0:

• Added EXPORT_PATH configuration
• Interactive mode with AskUserQuestion
• Direct mode for scripting/automation
• File paths updated to claude_usage/

Implementation

Configuration functionality is implemented in scripts/config.sh:

# Get skill directory
SKILL_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"

# Execute config script with arguments
bash "$SKILL_DIR/scripts/config.sh" "$@"

Script: skills/config/scripts/config.sh (150 lines)

Features:

• Interactive mode detection (no arguments)
• Direct mode with key=value parsing
• Configuration file validation
• Value validation with error messages
• Change detection and reporting
• Behavior explanation for each setting

See scripts/config.sh for full implementation details.