The Agent Skills Directory

[SAFE]: No security threats were identified across any analyzed categories.
[PROMPT_INJECTION]: The skill uses clear, instructional language to guide the agent in its task. It does not contain any attempts to bypass safety filters or override system instructions.
[DATA_EXFILTRATION]: The skill does not access sensitive local files or perform network requests. It operates only on the provided input strings.
[REMOTE_CODE_EXECUTION]: There are no patterns indicating the download or execution of external scripts or code.
[INDIRECT_PROMPT_INJECTION]: The skill processes external curriculum data which represents an ingestion surface. Ingestion points: reference_framework_band_tagged and comparison_frameworks_band_tagged in SKILL.md. Boundary markers: The prompt explicitly instructs the agent to quote content verbatim and attribute it within structured Markdown tables. Capability inventory: The skill has no capabilities involving subprocess calls, file-writes, or network operations. Sanitization: No explicit sanitization or filtering is described. Given the absence of dangerous tools or shell access, the surface is considered safe.

curriculum-crosswalk