The Agent Skills Directory

[COMMAND_EXECUTION]: Executes inline JavaScript via node -e to perform file system operations, such as calculating version strings and recursively removing cache directories (f.rmSync).
[COMMAND_EXECUTION]: Uses shell commands like rm -f and rm -rf to delete legacy scripts and directories (agents, commands, skills) during the automated fix process.
[EXTERNAL_DOWNLOADS]: Fetches remote content from a GitHub repository using WebFetch to update the local CLAUDE.md configuration file.
[DATA_EXFILTRATION]: Accesses and reads sensitive local configuration files, specifically ~/.claude/settings.json, to detect legacy hook configurations. While the data is used for local diagnostics, accessing configuration files is a prerequisite for potential exfiltration.
[PROMPT_INJECTION]: The skill ingests and processes content from user-controlled files like CLAUDE.md and settings.json. These ingestion points represent an indirect prompt injection surface where malicious instructions could be embedded to influence agent behavior.
Ingestion points: ~/.claude/CLAUDE.md, ./.claude/settings.json, ~/.claude/settings.json
Boundary markers: None detected in the provided scripts.
Capability inventory: File deletion (rm -rf), dynamic execution (node -e), and network fetching (WebFetch).
Sanitization: No explicit sanitization or validation of the ingested markdown or JSON content is performed before processing.

omc-doctor