The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from external sources which could contain malicious instructions.
Ingestion points: The skill accepts URLs, local PDF/image/audio files, and YouTube links as input (defined in SKILL.md).
Boundary markers: No explicit boundary markers or 'ignore' instructions are provided in the skill body to prevent the model from obeying instructions embedded in the summarized content.
Capability inventory: The skill executes the summarize binary, which may perform network requests and file reads.
Sanitization: No sanitization of input data is specified in the instruction set.
[COMMAND_EXECUTION]: The skill's primary functionality is to execute the summarize CLI tool with user-provided arguments.
[EXTERNAL_DOWNLOADS]: The skill documentation includes instructions to install the summarize binary from a third-party Homebrew tap (steipete/tap/summarize).

summarize