The Agent Skills Directory

[SAFE]: The skill consists of markdown content that provides persona framing, educational analogies, and philosophical arguments. It does not contain any scripts, binary executables, or automated commands that could compromise the host system.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to process user input through a simulated persona, which creates a natural surface for indirect prompt injection. However, the skill lacks integration with sensitive tools or system capabilities that would allow this surface to be exploited for harm.
Ingestion points: User requests defined in the interaction workflow in SKILL.md.
Boundary markers: The instructions explicitly require the agent to identify its responses as a simulation and to distinguish between facts and stylistic reconstructions.
Capability inventory: The skill does not utilize tools for file system modification, environment variable access, or unauthorized network operations.
Sanitization: There is no specific input validation or sanitization logic, as the skill relies on the agent's core safety protocols.

yann-lecun-filosofia