The Agent Skills Directory

[DYNAMIC_EXECUTION]: The skill instructs the agent to read and follow instructions from a file located at ~/.maestro/workflows/specs-add.md. This dynamic loading of logic from a local file outside the skill's own directory means the actual behavior is determined by external content that cannot be statically verified.\n- [INDIRECT_PROMPT_INJECTION]: The skill is designed to accept natural language 'intents' from users and pass them to a secondary workflow.\n
Ingestion points: User input is captured via the $ARGUMENTS variable in the dispatch block.\n
Boundary markers: There are no boundary markers or instructions to the model to ignore embedded commands within the user-supplied string.\n
Capability inventory: The skill has access to powerful tools including Bash, Write, and Edit, which could be abused if malicious instructions are processed.\n
Sanitization: There is no evident sanitization, validation, or filtering of the $ARGUMENTS content before it is passed to the workflow execution.

maestro-spec