The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to run the summarize binary for extracting and processing text content.
[EXTERNAL_DOWNLOADS]: The skill directs the agent to install a binary from a third-party Homebrew repository (steipete/tap/summarize), which is an unverifiable external dependency.
[INDIRECT_PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it processes untrusted data from the internet.
Ingestion points: The tool takes arbitrary URLs, YouTube links, and local file paths as input.
Boundary markers: No delimiters or safety instructions are used to distinguish between data and potential instructions within the source content.
Capability inventory: The underlying tool can read local files and perform network requests to external APIs.
Sanitization: The skill does not implement sanitization or validation for the ingested content.
[DATA_EXFILTRATION]: By design, the skill sends extracted content from local files or URLs to external LLM providers (such as Google, OpenAI, Anthropic, and xAI), which could lead to the exposure of sensitive data to these third parties.

summarize