cnki-parse-results
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill extracts various text fields (paper titles, authors, journals) from the
kns.cnki.netwebsite and incorporates them directly into the agent's context for presentation. - Ingestion points: Data is ingested through
mcp__chrome-devtools__evaluate_scriptandmcp__chrome-devtools__take_snapshotinSKILL.md. - Boundary markers: No explicit delimiters or instructions are used to prevent the agent from interpreting extracted content as new commands.
- Capability inventory: The agent context has access to browser automation tools (
mcp__chrome-devtools), which could be misused if a malicious search result successfully triggers an injection. - Sanitization: The skill performs basic whitespace trimming but does not sanitize or filter the content for prompt injection patterns.
- [COMMAND_EXECUTION]: The skill utilizes the
mcp__chrome-devtools__evaluate_scripttool to execute a JavaScript function within the browser's context. The executed code is a static parser for DOM elements, which is the intended functional purpose of the skill.
Audit Metadata