The Agent Skills Directory

[SAFE]: No security issues detected. The skill contains natural language instructions and configuration metadata without any executable scripts, shell commands, or network operations.
[INDIRECT_PROMPT_INJECTION]: The skill defines a workflow for ingesting and analyzing external references (code, websites, documentation), which is an inherent attack surface for indirect prompt injection.
Ingestion points: SKILL.md instructs the agent to catalog and inspect "Source code, library, existing component, screenshot, website, video, diagram, doc, issue, previous implementation, or product behavior."
Boundary markers: The instructions do not provide explicit delimiters or warnings to the agent to treat external content as data rather than instructions.
Capability inventory: The skill does not request special tools (allowed-tools is empty/missing), relying on the agent's default capabilities for file reading and web browsing.
Sanitization: No sanitization or validation of the ingested reference material is specified in the workflow.

reference-alignment