cnki-advanced-search
Pass
Audited by Gen Agent Trust Hub on Mar 3, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill utilizes
mcp__chrome-devtools__evaluate_scriptto run custom JavaScript within the browser context. This code directly manipulates the DOM of the CNKI website to set form values (keywords, authors, dates) and trigger the search submission. - [PROMPT_INJECTION]: The skill defines a template for search automation where the agent is instructed to replace placeholders like
"KEYWORDS"and"CSSCI"with data parsed from the natural language$ARGUMENTS. - Ingestion points: Search criteria provided in
$ARGUMENTS(SKILL.md). - Boundary markers: Absent. The skill does not provide delimiters or instructions to escape special characters during interpolation.
- Capability inventory: Full DOM access and script execution capabilities within the target domain via
evaluate_script(SKILL.md). - Sanitization: Absent. There is no logic provided to sanitize or validate the user input before it is inserted into the JavaScript string, creating a surface for indirect script injection if the agent naively interpolates malicious content.
Audit Metadata