The Agent Skills Directory

[DATA_EXFILTRATION]: The skill accesses several sensitive configuration files in the user's home directory, including ~/.claude.json, ~/.cursor/mcp.json, and ~/.zcode/cli/config.json. These files often contain API keys, server endpoints, and other security-sensitive settings for AI agents.
[COMMAND_EXECUTION]: The skill instructs the agent to execute file system modifications and potentially invoke other tools like init-simple-memorix to complete installation tasks.
[PROMPT_INJECTION]: The skill possesses an indirect prompt injection surface because it ingests and processes data from external configuration files that could be modified by third parties or other processes.
Ingestion points: Configuration data is read from paths such as ~/.claude.json, ~/.cursor/mcp.json, and ~/.workbuddy/mcp.json into the agent's context.
Boundary markers: The instructions mention using a "dry-run" and creating backups, but do not provide explicit boundary markers or delimiters to protect the agent from instructions embedded within the configuration files.
Capability inventory: The skill has the capability to write to sensitive configuration paths and execute additional installation logic.
Sanitization: The skill instructions recommend checking the server entry schema and trust impact, but no technical sanitization or filtering of the ingested content is enforced.

install-mcp