Security audit
Security checks for vulnerabilities and agentic risk
The skill is a disclosed financial-advisory helper that asks for startup information to produce analysis and documents, with sensitive-data handling notes but no hidden execution, exfiltration, persistence, or destructive behavior in the inspected artifact.
Before installing, treat any company plans, financial statements, founder biographies, investor lists, and PDFs as confidential. Use redacted or example data when possible, review any generated reports before sharing, and only run the optional system package installation commands if you understand the packages being installed.
Commands invoke sudo or root privileges. Verify this elevated access is necessary and justified.
brew install tesseract ghostscript poppler
sudo apt-get install tesseract-ocr poppler-utils ghostscript
### Basic Usage
This markdown file documents features that analyze startup financials and generate investor matching and outreach recommendations, which implies processing sensitive business and contact data. While the README includes financial-advice disclaimers, it does not warn users about privacy, confidentiality, or careful handling of uploaded company documents and investor data.
The README explicitly promotes parsing financial PDFs, OCR of scanned business plans, and generation of report files, all of which can affect user data, privacy, and local filesystems. There is no adjacent warning that these operations may extract sensitive information from documents or create output artifacts on disk.
The activation criteria are broad enough to match routine business, startup, investing, and planning questions, which increases the chance of unintended invocation. In a finance-oriented skill that solicits detailed company, team, and financial data, accidental activation can lead users to disclose sensitive information in contexts where they did not expect a specialized fundraising workflow or heightened data collection.
The skill instructs the agent to collect extensive sensitive business information, including revenue, expenses, runway, founder backgrounds, fundraising targets, and prior financing, without an upfront privacy warning or minimization guidance. Even if not overtly malicious, this creates unnecessary exposure of confidential startup and personal data, especially because the skill is designed to gather it conversationally over multiple exchanges.
At L479 the skill says it "can't read files," framing file access as unavailable, but elsewhere the documented behavior repeatedly offers to save generated outputs as files (L235, L296, L428-L431, L513). This is an intent-level contradiction in the skill documentation about filesystem capabilities, which could mislead users about what file operations the skill can perform.
No suspicious patterns detected.