weekly-competitor-tracker

Pass

Audited by Gen Agent Trust Hub on May 15, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill operates exclusively on local markdown files provided by the user. It performs text parsing and diffing using natural language instructions without the use of high-risk tools such as shell execution or network access.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted content from external markdown files, which constitutes a potential surface for indirect prompt injection. However, the lack of dangerous capabilities significantly limits the risk.
  • Ingestion points: Reads content from files in the competitors/ folder and competitors/snapshot/ subfolder (SKILL.md, Step 2 & 3).
  • Boundary markers: The instructions do not define specific delimiters or instructions to ignore embedded commands within the files.
  • Capability inventory: File read and write access to the local filesystem for generating the .md report. No network, shell, or code execution tools are utilized.
  • Sanitization: No explicit sanitization or validation of the input file content is defined in the instructions.
Audit Metadata
Risk Level
SAFE
Analyzed
May 15, 2026, 09:40 AM
Security Audit — agent-trust-hub — weekly-competitor-tracker