20206 02 10 Clawhub Summarize 1.0.0

PassAudited by ClawScan on May 1, 2026.

Overview

This appears to be a straightforward summarization CLI skill, with normal but important reliance on an external Homebrew package, external model providers, and optional API keys.

This skill is reasonable for its stated purpose. Before installing, verify that you trust the `summarize` Homebrew package and be careful about which API keys and local files or private URLs you use with it.

Findings (3)

Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.

What this means

Installing the skill may install and run a third-party CLI package on the user's machine.

Why it was flagged

The skill relies on an external Homebrew formula to provide the `summarize` binary. This is expected for a CLI-based skill, but it means installation trust depends on that package source.

Skill content
brew | formula: steipete/tap/summarize | creates binaries: summarize
Recommendation

Install only if you trust the Homebrew tap and project; review the formula/source if provenance is important to you.

What this means

If configured, the CLI can use the user's provider accounts for model calls, web extraction, or YouTube fallback processing.

Why it was flagged

The skill documents use of provider and service credentials. This is purpose-aligned for summarization, but those credentials grant access to third-party accounts or quotas.

Skill content
Set the API key for your chosen provider: - OpenAI: `OPENAI_API_KEY` - Anthropic: `ANTHROPIC_API_KEY` ... Optional services: - `FIRECRAWL_API_KEY` ... - `APIFY_API_TOKEN`
Recommendation

Use only the keys needed for your chosen provider, prefer restricted keys when available, and revoke or rotate keys if they are no longer needed.

What this means

Sensitive documents, URLs, transcripts, or media could be sent to external services as part of normal summarization use.

Why it was flagged

The examples combine local-file summarization with an external model selection. This is disclosed and expected, but it means file or URL content may be processed by the selected provider or fallback service.

Skill content
summarize "/path/to/file.pdf" --model google/gemini-3-flash-preview
Recommendation

Avoid using highly sensitive files or private URLs unless the selected provider's privacy and retention terms are acceptable.