skills/larksuite/cli/lark-doc/Gen Agent Trust Hub

lark-doc

Pass

Audited by Gen Agent Trust Hub on Oct 28, 2026

Risk Level: SAFE
Full Analysis
  • [DATA_EXPOSURE_AND_EXFILTRATION]: The skill interacts with Feishu/Lark document content and media resources. It implements robust security controls to prevent unauthorized data exfiltration or SSRF. For example, in references/lark-doc-fetch.md and references/lark-doc-resource-cover.md, the skill explicitly validates resource URLs, requiring HTTPS and forbidding private, loopback, or multicast IP addresses. It also limits redirect following to prevent potential bypasses.
  • [INDIRECT_PROMPT_INJECTION]: The skill ingests data from external Lark documents, which could contain malicious instructions.
  • Ingestion points: Document content and comments are read via references/lark-doc-fetch.md.
  • Boundary markers: The skill uses structured XML and Markdown formats and enforces strict "Genre Contracts" to maintain content integrity.
  • Capability inventory: Document creation and modification (references/lark-doc-create.md, references/lark-doc-update.md) and filesystem writes via media downloads (references/lark-doc-media-download.md).
  • Sanitization: The skill mandates XML escaping (references/lark-doc-xml.md) and uses a specialized script for parsing and validating document profiles and resources (references/lark-doc-script.md).
  • [COMMAND_EXECUTION]: The skill uses platform-specific command-line utilities (osascript, PowerShell, xclip, wl-paste, xsel) exclusively for clipboard operations as part of its documented functionality to insert media. All other operations are confined to the authorized lark-cli tool.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 28, 2026, 08:48 PM