mtga-draft-helper

Warn

Audited by Gen Agent Trust Hub on Feb 28, 2026

Risk Level: MEDIUMCOMMAND_EXECUTIONEXTERNAL_DOWNLOADSDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill instructs the agent to run a background Python script (live_watcher.py) and PowerShell commands to monitor the game log in real-time as described in references/live-draft-tracking.md.
  • [DATA_EXFILTRATION]: The skill reads the MTG Arena Player.log file from the user's local application data directory to extract draft state information as detailed in references/arena-log-parsing.md.
  • [EXTERNAL_DOWNLOADS]: The skill fetches card performance statistics and images from community services including 17lands.com and scryfall.com.
  • [PROMPT_INJECTION]: The skill is subject to indirect prompt injection because it processes data from local files and external APIs without explicit sanitization.
  • Ingestion points: Draft data is ingested from the local Player.log file (referenced in arena-log-parsing.md) and the 17Lands API.
  • Boundary markers: There are no explicit markers or instructions to disregard embedded commands used when processing log or API data.
  • Capability inventory: The skill possesses the ability to execute system commands, run background scripts, and read local files.
  • Sanitization: Data from the log and API are parsed for specific keys but not sanitized for potential malicious instructions before being used in suggestion prompts.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Feb 28, 2026, 09:31 PM