summarize
Pass
Audited by Gen Agent Trust Hub on Apr 24, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill installs the
summarizeutility from the author's custom Homebrew tap (steipete/tap/summarize). - [COMMAND_EXECUTION]: The agent executes the
summarizecommand-line tool to process user-provided inputs such as URLs, YouTube links, and local PDF files. - [DATA_EXFILTRATION]: The tool transmits content to external well-known AI services (OpenAI, Anthropic, Google, xAI) to generate summaries. This is the intended primary function of the tool and requires the user to provide their own API keys.
- [PROMPT_INJECTION]: The skill is subject to indirect prompt injection risks as it processes content from untrusted external sources (web pages and video transcripts) which could contain malicious instructions designed to manipulate the agent's behavior. This is an inherent risk common to all summarization tasks.
Audit Metadata