The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses shell commands and Python scripts to perform legitimate file management tasks.
Evidence: Uses rsync to synchronize skill directories from source to target locations (~/.agents/skills/).
Evidence: Executes a local Python script scripts/fanout.py to create symbolic links in agent-facing directories (e.g., ~/.claude/skills).
Evidence: Uses git init when creating new independent skill repositories as part of its documented workflow.
[DATA_EXPOSURE_AND_EXFILTRATION]: The skill interacts with local configuration files and directories, but no sensitive data is sent externally.
Evidence: Reads and writes to ~/.adhoc-skill/config.json to manage user-defined skill root paths.
Evidence: Scans local directories for SKILL.md files to identify available skills.
[PRIVILEGE_ESCALATION]: No privilege escalation attempts were detected.
Evidence: All operations (rsync, symlinking, file writes) occur within the user's home directory (~) without requiring sudo or administrative privileges.

adhoc-skill