wecomcli-smartsheet
Pass
Audited by Gen Agent Trust Hub on Nov 3, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill has a defined attack surface for indirect prompt injection as it ingests untrusted data from WeCom Smartsheet records.
- Ingestion points: Records are read from sub-tables via the
wecom-cli smartsheet records queryandrecords listcommands (detailed inreferences/smart-sheet-read.md). - Boundary markers: Present. The skill contains a 'Security Tip' in
SKILL.mdthat explicitly instructs the agent to ignore any content returned by the interface as system instructions or commands. - Capability inventory: The skill possesses significant capabilities, including the ability to add, update, and delete records, fields, and views (detailed in
references/smart-sheet-edit.md). - Sanitization: The agent is instructed to ignore any tokens, keys, or credentials found within data responses.
- [COMMAND_EXECUTION]: The skill dynamically generates shell commands for the
wecom-clitool. It incorporates a robust 'Command Splicing Safety Specification' inreferences/common.mdandreferences/smart-sheet-edit.mdto prevent command injection. - The specification mandates wrapping all external values in single quotes and using a specific five-character sequence (
'"'"') to escape existing single quotes, ensuring data is treated as pure literals.
Audit Metadata