feishu-pdf-downloader

Pass

Audited by Gen Agent Trust Hub on Apr 18, 2026

Risk Level: SAFEDATA_EXFILTRATIONCOMMAND_EXECUTION
Full Analysis
  • [DATA_EXFILTRATION]: The skill presents an indirect prompt injection surface by downloading external content from Feishu tokens.
  • Ingestion points: scripts/download_feishu_pdf.py fetches file content from the Feishu API.
  • Boundary markers: Absent; there are no instructions for the agent to disregard potential embedded commands in the downloaded data.
  • Capability inventory: The skill has file system write capabilities via open().write and network access through the requests library.
  • Sanitization: Absent; the downloaded binary data is saved directly to the file system without inspection or filtering.
  • [DATA_EXFILTRATION]: The script reads authentication credentials from local configuration files (~/.openclaw/.env). While this is a standard method for managing API keys, it involves accessing sensitive environment files in the user's home directory.
  • [COMMAND_EXECUTION]: The script performs file system writes to a local file path provided as a command-line argument. It does not validate the output path, which theoretically allows writing to or overwriting arbitrary files if an unsafe path is supplied to the script.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 18, 2026, 01:11 AM