tencent-docs
Pass
Audited by Gen Agent Trust Hub on May 19, 2026
Risk Level: SAFEREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
- [REMOTE_CODE_EXECUTION]: The
sheet.operation_sheettool (defined insheet/api/operation-api.md) allows the agent to generate and execute JavaScript code on the Tencent Docs platform. This is a high-capability feature that executes AI-generated logic against remote spreadsheet APIs. - [COMMAND_EXECUTION]: The skill uses local shell scripts (
setup.sh,import_file.sh) and a Node.js utility (generate_slide.js) to manage tokens, perform file uploads, and handle asynchronous PPT generation. These scripts interact with the system environment and themcporterCLI tool. - [EXTERNAL_DOWNLOADS]: The
setup.shscript performs a global installation of themcporterNode.js package. The skill also facilitates authentication by fetching tokens fromdocs.qq.comand uploading files to cloud storage URLs. - [PROMPT_INJECTION]: The skill has multiple prompt injection surfaces. First, the
js-script-rule.mdfile contains defensive instructions against behavior overrides. Second, thescrape_urlworkflow ingests untrusted external content from the web into the agent's context without explicit sanitization or boundary markers (Category 8), potentially allowing indirect injection from malicious websites.
Audit Metadata