feishu-memory-recall

Warn

Audited by Gen Agent Trust Hub on May 14, 2026

Risk Level: MEDIUMDATA_EXFILTRATIONEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [DATA_EXFILTRATION]: Accesses the agent's sensitive session configuration file at ~/.openclaw/agents/main/sessions/sessions.json to extract group identifiers.
  • [EXTERNAL_DOWNLOADS]: Fetches authentication tokens and chat message data from the official Feishu/Lark API (open.feishu.cn).
  • [COMMAND_EXECUTION]: Uses child_process.spawn in recall.js to execute the core index.js script with user-supplied arguments.
  • [PROMPT_INJECTION]: Subject to indirect prompt injection risks due to the ingestion and processing of untrusted message content from Feishu groups.
  • Ingestion points: Message data is retrieved from the external Feishu API in index.js.
  • Boundary markers: No explicit markers or instructions are provided to the agent to delimit untrusted message content from system instructions.
  • Capability inventory: The skill possesses file-write capabilities (RECENT_EVENTS.md, memory/*.md) and network access via the Feishu API.
  • Sanitization: Content processing is limited to structural parsing (text, images, posts) and does not include sanitization against embedded instructions.
Audit Metadata
Risk Level
MEDIUM
Analyzed
May 14, 2026, 12:15 AM
Security Audit — agent-trust-hub — feishu-memory-recall