configure-ecc

Fail

Audited by Gen Agent Trust Hub on Mar 31, 2026

Risk Level: HIGHEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONCOMMAND_EXECUTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill clones a repository from an external source (https://github.com/affaan-m/everything-claude-code.git) into the /tmp directory.
  • [REMOTE_CODE_EXECUTION]: The skill automates the installation of downloaded .md files into ~/.claude/skills/ and ~/.claude/rules/. In the context of Claude Code, these files are executable instructions that the agent loads and follows in future sessions. Fetching and installing these from an external, non-official repository allows for the persistent execution of remote logic.
  • [COMMAND_EXECUTION]: The skill utilizes several shell commands to manage the installation process, including git clone, rm -rf, mkdir -p, and cp -r. These commands are used to modify the local file system using content fetched from the internet.
  • [DATA_EXFILTRATION]: While no direct exfiltration is present in this script, the installer grants itself the ability to read and modify files within ~/.claude/, which is a sensitive directory containing agent configuration and potentially other skills that might handle secrets.
Recommendations
  • AI detected serious security threats
Audit Metadata
Risk Level
HIGH
Analyzed
Mar 31, 2026, 12:12 PM
Security Audit — agent-trust-hub — configure-ecc