DeepRead BYOK

Security checks across malware telemetry and agentic risk

Overview

This instruction-only DeepRead BYOK skill is coherent and disclosed, but users should understand it involves API keys, provider billing, document uploads, and an optional separate setup skill.

This skill appears safe to install as an instruction-only guide, but treat it as a credentialed third-party integration: use limited API keys where possible, set budget limits, review DeepRead and provider data policies, and separately review the optional OAuth setup skill before installing it.

VirusTotal

VirusTotal findings are pending for this skill version.

View on VirusTotal

Risk analysis

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.

#
ASI03: Identity and Privilege Abuse
Medium
What this means

Your provider account may be billed for document processing, and the connected key could have significant account authority depending on how it is scoped.

Why it was flagged

The skill requires use of a DeepRead API key and asks the user to connect a provider API key, which gives DeepRead/provider-side workflows billing and processing authority under the user's account.

Skill content
Use your own OpenAI, Google, or OpenRouter API key for all DeepRead document processing. Your key, your billing, zero DeepRead LLM costs.
Recommendation

Use a restricted provider key if available, set spending limits, monitor provider usage, and delete or rotate the key if you stop using BYOK.

#
ASI07: Insecure Inter-Agent Communication
Low
What this means

Documents you process may leave your local environment and be handled by DeepRead and the selected model provider.

Why it was flagged

The documented workflow sends local documents to DeepRead's API for processing, and BYOK routes LLM usage through the user's provider account.

Skill content
curl -X POST https://api.deepread.tech/v1/process \
  -H "X-API-Key: $DR_API_KEY" \
  -F "file=@document.pdf"
Recommendation

Only process documents you are comfortable sending to DeepRead and the chosen provider, and review their data handling policies for sensitive or regulated files.

#
ASI04: Agentic Supply Chain Vulnerabilities
Info
What this means

Installing the separate setup skill may grant additional authentication-related capabilities that were not reviewed here.

Why it was flagged

The reviewed skill is instruction-only, but it points users to install a separate setup skill that is not included in the provided artifacts.

Skill content
For automated agent setup with OAuth device flow, install the dedicated skill:

```bash
clawhub install uday390/deepread-agent-setup
```
Recommendation

Review the separate deepread-agent-setup skill and its requested permissions before installing it.