claude-skills-troubleshooting
Pass
Audited by Gen Agent Trust Hub on May 12, 2026
Risk Level: SAFEDATA_EXFILTRATIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
- [DATA_EXFILTRATION]: The skill accesses internal configuration and registry files within the
~/.claude/directory, such assettings.jsonandinstalled_plugins.json. This data exposure is localized to the agent's environment and is required for the skill's primary purpose of diagnosing plugin installation and activation states. - [COMMAND_EXECUTION]: Executes bundled Python scripts (
diagnose_plugins.pyandenable_all_plugins.py) and shell commands to automate diagnostics and update user configuration settings upon verification. - [EXTERNAL_DOWNLOADS]: Contains references and links to official documentation and issue trackers hosted on
github.com/anthropicsandclaude.com. These links point to trusted sources and are used to provide context for known architectural issues within the Claude Code platform.
Audit Metadata