skills/edwinhu/workflows/cite-check/Gen Agent Trust Hub

cite-check

Pass

Audited by Gen Agent Trust Hub on Jun 22, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [COMMAND_EXECUTION]: The skill spawns external processes including rclone, readwise, and python3. These tools are used for document materialization from cloud services and PDF text extraction. Command arguments are passed via arrays in spawnSync, which prevents shell injection vulnerabilities.
  • [EXTERNAL_DOWNLOADS]: The skill fetches document content and metadata from well-known services including Google Drive (via rclone) and Readwise. These operations are essential to the skill's primary purpose of verifying academic sources.
  • [PROMPT_INJECTION]: The skill constructs prompts for the Gemini API that incorporate claim text extracted from local Markdown drafts. While this presents a surface for indirect prompt injection, the prompts are structured with specific instructions to the model to minimize the risk of behavior override.
  • [SAFE]: The skill follows secure practices for handling API keys (recommending environment variables) and uses established libraries and standard CLI tools for its operations. No malicious patterns such as obfuscation or unauthorized data exfiltration were detected.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 22, 2026, 06:30 AM
Security Audit — agent-trust-hub — cite-check