A desktop application for exploring, analyzing, and debugging Claude Code skills. Built with Tauri, React, and TypeScript, it helps operations teams identify and resolve issues in their AI workflows.
git clone https://github.com/SpillwaveSolutions/skills_viewer.gitThe skills_viewer is an essential tool designed for developers and AI practitioners who need to efficiently look at and debug the Claude Code Skills they are using. This intermediate-level automation skill streamlines the debugging process, allowing users to quickly identify issues within their AI automation workflows. By providing a clear view of the skills in action, it enhances the overall development experience and ensures smoother integration of AI agent skills into projects. One of the key benefits of the skills_viewer is its ability to save time during the debugging phase. While exact time savings are currently unknown, the skill's intuitive interface and functionality can significantly reduce the hours spent troubleshooting issues, enabling teams to focus more on development and innovation. This is especially valuable in fast-paced environments where efficiency is paramount. The skills_viewer is particularly suited for developers, product managers, and AI practitioners who are involved in creating or managing AI automation solutions. It is a valuable addition to any team looking to optimize their workflow automation processes. For example, a developer working on a data engineering project can utilize the skills_viewer to quickly debug any issues with the Claude Code Skills related to data processing, ensuring a more reliable and efficient system. Implementation of the skills_viewer is straightforward, requiring approximately 30 minutes to set up. Given its intermediate complexity, users should have a basic understanding of Claude Code and AI automation principles. This skill fits seamlessly into AI-first workflows, providing a critical layer of oversight and control that enhances the reliability of AI agent skills. By integrating skills_viewer into your development toolkit, you can ensure that your AI automation projects run smoothly and effectively.
[{"step":"Access the Sortd Skills Viewer","action":"Open Sortd for Gmail and navigate to the 'AI Features' dashboard. Click on 'Skills Viewer' in the left sidebar to load the tool.","tip":"Ensure you have admin permissions for the Sortd workspace to view and edit skill configurations."},{"step":"Select the Skill to Debug","action":"In the Skills Viewer, use the search bar to find the skill you want to analyze (e.g., 'AI-Powered Lead Prioritization'). Click on the skill to open its detailed view.","tip":"Filter skills by 'Status' (e.g., 'Failed' or 'Warning') to quickly identify problematic skills."},{"step":"Review Execution Logs and Configuration","action":"In the skill's detail view, review the 'Execution Logs' tab to identify patterns in failures (e.g., time of day, error messages). Then, switch to the 'Configuration' tab to check parameters like triggers, dependencies, and thresholds.","tip":"Use the 'Export Logs' button to download detailed logs for further analysis in a spreadsheet tool."},{"step":"Apply Recommended Fixes","action":"Based on the analysis, update the skill's configuration or code directly in the Skills Viewer. For example, adjust the 'priority_threshold' parameter or upgrade dependencies. Save changes and trigger a test execution.","tip":"Use Sortd's 'Test Skill' button to validate changes before deploying to production."},{"step":"Monitor Results and Iterate","action":"After deploying fixes, monitor the skill's performance in the Skills Viewer for 48 hours. Check the 'AI Status Tracker' for improvements in reliability and the 'Execution Logs' for reduced failures.","tip":"Set up a Slack or email alert in Sortd's 'Notifications' settings to be notified of any new failures."}]
No install command available. Check the GitHub repository for manual installation instructions.
git clone https://github.com/SpillwaveSolutions/skills_viewerCopy the install command above and run it in your terminal.
Launch Claude Code, Cursor, or your preferred AI coding agent.
Use the prompt template or examples below to test the skill.
Adapt the skill to your specific use case and workflow.
Use the Sortd skills_viewer to debug and analyze the AI skill '[SKILL_NAME]' in the Sortd workspace. Review the skill's configuration, dependencies, and recent execution logs to identify: 1) Potential issues causing intermittent failures, 2) Missing or misconfigured [PLACEHOLDER: parameters like triggers, permissions, or integrations], 3) Performance bottlenecks in [PLACEHOLDER: specific workflow steps]. Provide actionable recommendations to resolve these issues and optimize the skill for reliability. Reference the Sortd AI features documentation for status tracking, urgency detection, and complaint handling if relevant.
### Sortd Skills Viewer Debug Report: 'AI-Powered Lead Prioritization Skill' **Skill Overview:** The 'AI-Powered Lead Prioritization' skill is designed to automatically triage incoming sales leads from Gmail using Sortd's kanban boards and Claude Code. It assigns urgency scores based on email content (e.g., keywords like 'urgent' or 'ASAP') and moves high-priority leads to the 'Hot Leads' column. The skill has been experiencing intermittent failures over the past 48 hours, with 32% of executions failing to update the board or log results. **Key Findings from Sortd Skills Viewer:** 1. **Configuration Issues:** - The skill's trigger (Gmail label 'Incoming-Leads') is correctly configured, but the 'priority_threshold' parameter is set to a static value of 0.7. This is too high for our dataset, where only 12% of leads meet this threshold. The average urgency score in the last 7 days is 0.45, causing most leads to be ignored. - The 'Claude Code' dependency (v1.2.3) is outdated. The latest version (v1.4.0) includes fixes for API rate limits that may be causing timeouts. 2. **Execution Logs Analysis:** - 18 of the 25 failed executions (72%) occurred between 2-4 PM ET, coinciding with peak Gmail API usage. The logs show repeated '429 Too Many Requests' errors when querying Sortd's kanban board API. - The remaining failures were due to missing 'lead_source' metadata in 7 leads, causing the skill to crash when attempting to categorize them. 3. **Performance Bottlenecks:** - The skill's 'analyze_email' function takes an average of 12 seconds to process each lead, with 80% of the time spent on Claude Code's NLP analysis. This delay causes Sortd's AI status tracker to mark the skill as 'stalled' after 10 seconds of inactivity. **Recommended Fixes:** - **Adjust Priority Threshold:** Lower the 'priority_threshold' to 0.35 to align with actual urgency scores. This should increase the number of prioritized leads from 12% to ~45%. - **Update Dependencies:** Upgrade Claude Code to v1.4.0 and add a 5-second retry delay for API rate limits. - **Handle Missing Metadata:** Add a fallback mechanism to tag leads with 'lead_source: unknown' when metadata is missing, preventing crashes. - **Optimize NLP Analysis:** Split the 'analyze_email' function into two steps: 1) Quick keyword scan (1-2 seconds) to flag urgent leads, 2) Full NLP analysis only for leads passing the initial scan. This reduces average processing time to 6 seconds. **Next Steps:** 1. Update the skill configuration in Sortd's 'AI Features' dashboard under 'Lead Prioritization Settings'. 2. Deploy the updated Claude Code dependency via Sortd's 'Skills Manager'. 3. Monitor the skill's execution logs for 48 hours to verify improvements. The Sortd AI status tracker should now show 'Running' status consistently, with no stalled executions. **Impact:** - Expected reduction in failed executions from 32% to <5%. - Faster lead triage, with 80% of leads processed within 6 seconds (vs. 12 seconds previously). - Improved team productivity, as sales reps will no longer miss urgent leads due to skill failures.
AI assistant built for thoughtful, nuanced conversation
Get more done every day with Microsoft Teams – powered by AI
Automate security compliance and monitor real-time security posture seamlessly.
Automate your spreadsheet tasks with AI power
Agentic AI Workflow platform
Connected workspace for docs, wikis, and projects
Take a free 3-minute scan and get personalized AI skill recommendations.
Take free scan