The Agent Skills Directory

Prompt Injection (SAFE): The instructions follow a pedagogical structure (Socratic, Feynman, Problem-Based) and do not contain attempts to bypass safety filters, extract system prompts, or override agent constraints.
Data Exposure & Exfiltration (SAFE): The skill uses Read, finder, and Grep to access the local codebase. This is intended for the purpose of explaining project-specific code to the user. No credentials, sensitive system paths, or unauthorized network exfiltration patterns were detected.
External Downloads & RCE (SAFE): The skill does not download external scripts or packages. All teaching methodologies are stored locally in the cookbook/ directory.
Indirect Prompt Injection (LOW): The skill has an ingestion surface for untrusted data via web_search and read_web_page.
Ingestion points: Data retrieved from external URLs via read_web_page and local files via Read.
Boundary markers: Absent; the skill does not explicitly define delimiters for untrusted data.
Capability inventory: The skill is limited to Read, web_search, read_web_page, finder, and Grep. It lacks high-risk capabilities such as file-writing or arbitrary command execution.
Sanitization: Absent; the skill processes retrieved content directly for synthesis into teaching material.
Obfuscation (SAFE): No encoded strings, hidden characters, or homoglyphs were found in the skill files.

teacher