@tank/cooking-mastery

Pass

Audited by Gen Agent Trust Hub on May 31, 2026

Risk Level: SAFEPROMPT_INJECTIONDATA_EXFILTRATIONEXTERNAL_DOWNLOADS
Full Analysis
  • [PROMPT_INJECTION]: The skill processes untrusted data from external websites and social media platforms, creating a surface for indirect prompt injection.
  • Ingestion points: Scrapes recipe data, descriptions, and transcripts from YouTube, TikTok, Instagram, and recipe blogs as detailed in the video extraction and search references.
  • Boundary markers: There are no defined delimiters or instructions to ignore embedded commands within the extracted external data.
  • Capability inventory: The skill has permissions to write to local storage (~/.cooking-mastery/cookbook.json) and communicate with the Bring! API for list management.
  • Sanitization: Extraction logic focuses on structural parsing but does not include steps to sanitize or filter potential malicious instructions within the scraped text.
  • [DATA_EXFILTRATION]: The skill requests broad read access to the entire file system hierarchy (**/*) in its manifest, which is unnecessary for its stated purpose and could allow access to sensitive user files. It also transmits shopping list data to the Bring! service.
  • [EXTERNAL_DOWNLOADS]: The skill performs automated downloads and scraping from external sources, including TheMealDB API and major social media platforms like YouTube and TikTok, to retrieve recipe information.
Audit Metadata
Risk Level
SAFE
Analyzed
May 31, 2026, 11:04 PM
Security Audit — agent-trust-hub — @tank/cooking-mastery