The Agent Skills Directory

[COMMAND_EXECUTION]: The skill exposes an interface for arbitrary file modification using the --edit and --content parameters. This allows the agent to write any content to any file path accessible to the script, which could be used to overwrite sensitive configurations or inject malicious scripts if the agent is compromised or manipulated.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to search through and ingest codebase content, which is a potential surface for indirect prompt injection.
Ingestion points: Content from the codebase is retrieved and placed into the agent's context via the --search operation described in SKILL.md.
Boundary markers: There are no specific delimiters or instructions defined to isolate search results or prevent the agent from following instructions found within the code being searched.
Capability inventory: The skill uses Bash and Read tools, and provides a custom file-writing capability via the --edit flag.
Sanitization: The documentation does not describe any sanitization or filtering of the content retrieved from the file system before it is processed by the agent.

morph-search