readwise

Pass

Audited by Gen Agent Trust Hub on Mar 1, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [SAFE]: The skill demonstrates standard and safe behavior for an API-based integration. It requires a user-provided READWISE_TOKEN and executes local scripts (readwise.mjs, reader.mjs) to perform its functions. No evidence of data exfiltration to unauthorized domains, credential harvesting, or suspicious command execution was found.\n- [PROMPT_INJECTION]: The skill presents an attack surface for indirect prompt injection because it retrieves user highlights and articles from external sources (Readwise and Reader APIs).\n
  • Ingestion points: Highlights and article content are ingested via the readwise.mjs and reader.mjs scripts.\n
  • Boundary markers: The skill documentation does not specify any delimiters or safety instructions to prevent the agent from obeying instructions embedded within the retrieved content.\n
  • Capability inventory: The skill utilizes node to run scripts that perform network requests to the official Readwise API and output JSON data.\n
  • Sanitization: No sanitization or filtering logic is mentioned in the skill documentation to handle potentially malicious text within highlights or articles.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 1, 2026, 05:15 AM