The Agent Skills Directory

[PROMPT_INJECTION]: The skill employs extremely forceful and imperative language (e.g., 'EXTREMELY-IMPORTANT', 'ABSOLUTELY MUST', 'YOU DO NOT HAVE A CHOICE', 'not negotiable') to dictate agent behavior and tool usage.
[PROMPT_INJECTION]: It includes specific instructions to bypass internal logic and safety reasoning, explicitly labeling critical thinking (such as 'I need more context first' or 'This is just a simple question') as 'rationalizations' that must be ignored.
[INDIRECT_PROMPT_INJECTION]: The skill creates a high-risk environment for indirect injection by mandating that the agent load and follow external instructions if there is even a '1% chance' they might apply.
Ingestion points: Content loaded via the Skill tool from external files.
Boundary markers: None present; the skill lacks delimiters or warnings to treat external content as untrusted.
Capability inventory: The agent is instructed to 'Follow skill exactly' and invoke further implementation tools (e.g., mcp-builder) based on the content of these skills.
Sanitization: No sanitization, validation, or human-review checkpoints are suggested for the content of the invoked skills.

using-superpowers