jaspr-convert-html
Pass
Audited by Gen Agent Trust Hub on Jun 14, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill relies on the execution of the
jaspr convert-htmlcommand in the shell to process data. - [EXTERNAL_DOWNLOADS]: The tool includes a
--urlparameter that enables the agent to fetch content from remote web servers, which is then processed locally. - [PROMPT_INJECTION]: The skill presents an attack surface for indirect prompt injection (Category 8).
- Ingestion points: The skill ingests untrusted data from both local files (via
--file) and remote web pages (via--url). - Boundary markers: Absent. There are no instructions or delimiters provided to help the agent distinguish between the HTML data being processed and control instructions.
- Capability inventory: The skill uses shell command execution (
jaspr). - Sanitization: Absent. Rule 3 explicitly commands the agent to use the output "UNCHANGED" and "exactly as provided," which could force the agent to follow malicious instructions if they are embedded within the HTML source and preserved in the conversion output.
Audit Metadata