space-weread-analyzer

Pass

Audited by Gen Agent Trust Hub on May 20, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONCREDENTIALS_UNSAFEPROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill communicates with Tencent's official WeChat Reading API (i.weread.qq.com) to retrieve user notebooks, bookmarks, reviews, and reading statistics.
  • [COMMAND_EXECUTION]: Executes an internal Python script (fetch.py) to aggregate data and utilizes system-level commands (open, xdg-open, or start) to automatically launch the generated HTML report in the user's default browser.
  • [CREDENTIALS_UNSAFE]: The skill utilizes a WEREAD_API_KEY stored in environment variables to authenticate its requests to the WeChat Reading service.
  • [PROMPT_INJECTION]: The skill processes untrusted user-generated content in the form of book notes and highlights, creating an indirect prompt injection surface.
  • Ingestion points: User reading notes and highlights retrieved via the scripts/fetch.py script and stored in /tmp/space-weread-analyzer/data.json.
  • Boundary markers: None identified in the instructions for isolating ingested data during the analysis phase.
  • Capability inventory: File system access (reading data, writing HTML reports), directory creation (mkdir -p), and subprocess execution (python3, system open).
  • Sanitization: The skill instructions attempt to mitigate risks by requiring the agent to paraphrase insights and strictly limiting the length of direct quotes.
Audit Metadata
Risk Level
SAFE
Analyzed
May 20, 2026, 01:38 AM
Security Audit — agent-trust-hub — space-weread-analyzer