fsi-strip-profile

Pass

Audited by Gen Agent Trust Hub on Feb 25, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • Command Line Execution for File Conversion: The skill instructions involve executing system commands such as soffice and pdftoppm to convert PowerPoint files into images for visual review.
  • Evidence: The Workflow section in SKILL.md specifies the use of soffice --headless --convert-to pdf and pdftoppm -jpeg within a bash block.
  • Context: This behavior is directly associated with the skill's primary purpose of ensuring document quality and visual alignment before presenting to the user.
  • Indirect Prompt Injection Surface: The skill is designed to research and process data from external sources such as SEC filings, investor presentations, and news releases.
  • Ingestion points: SKILL.md (Research & Planning section) specifies data entry from BamSEC, SEC EDGAR, corporate websites, and news press releases.
  • Boundary markers: No specific delimiters or "ignore instructions" warnings are defined for the external data ingestion.
  • Capability inventory: The skill utilizes PptxGenJS and Python for file generation, and shell commands for image conversion.
  • Sanitization: There are no explicit instructions for sanitizing or filtering instructions that might be embedded in the external financial documents.
  • Context: While this represents an attack surface where malicious content in a processed document could attempt to influence the agent, it is a standard risk for agents performing research tasks and is managed by the underlying model's safety guardrails.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 25, 2026, 02:25 AM