The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests and processes user-provided profile files (person-dna.md) and content drafts to provide tailored diagnostic feedback. This creates a surface for indirect prompt injection where instructions hidden within these data files could attempt to influence the agent's behavior.
Ingestion points: The agent reads the file path provided by the user for the coach's DNA profile and analyzes the content of provided drafts.
Boundary markers: The instructions do not specify the use of clear delimiters or instructions to ignore potential commands embedded within the user-provided profile or draft content.
Capability inventory: The skill has the capability to execute a bundled validation script and access file contents to perform its primary function.
Sanitization: While the bundled script validates the structural integrity and activation status of the profile, it does not sanitize the text content for malicious prompt injection patterns.
[COMMAND_EXECUTION]: The skill relies on the execution of a bundled Python script (scripts/check_profile.py) to verify that the user's profile is active and correctly formatted before proceeding with the diagnosis. This script uses standard Python libraries and operates locally on the provided file path without performing network requests or writing to the file system.

yetong-review