The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes external, user-supplied text (the 'case') to perform its reasoning steps. If the input case contains instructions intended to override the agent's behavior, the agent might attempt to follow them.
Ingestion points: The skill accepts a <case to examine> as input in the README.md and SKILL.md instructions.
Boundary markers: The skill does not use specific delimiters or warnings to isolate the user-provided case from the reasoning instructions.
Capability inventory: The skill is restricted to text generation and does not utilize tools for file system access, network operations, or code execution.
Sanitization: There is no evidence of input validation or escaping for the processed case content.
[EXTERNAL_DOWNLOADS]: The skill's documentation includes instructions and links for downloading components from the vendor's repository.
The README.md provides an installation command using npx that targets https://github.com/synapseradio/ai-skills.
The README.md contains a direct link to a packaged skill file at https://github.com/synapseradio/ai-skills/raw/main/packaged/thinkies/generalize.skill.

generalize