The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes external task requirements and project code while possessing file-writing capabilities, which creates a vulnerability surface for instructions embedded in external data.
Ingestion points: The agent ingests and analyzes hardware-adjacent code, driver logic, and system constraints provided in the workspace (SKILL.md).
Boundary markers: There are no explicit delimiters or instructions provided to the agent to disregard instructions that might be embedded within the files it analyzes.
Capability inventory: The skill is configured with workspace-write permissions, allowing it to implement changes or recommendations directly to the filesystem.
Sanitization: The instructions do not include requirements to sanitize, escape, or validate content retrieved from the workspace before it is processed or used to generate new code.

embedded-systems