ajtbd-b2c-segments
Pass
Audited by Gen Agent Trust Hub on Mar 5, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill utilizes the
Bashtool to perform local filesystem operations, specifically runningls .ajtbd/andmkdir -p .ajtbdto manage a hidden directory for research persistence. - [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it is designed to analyze content from user-provided URLs without sufficient isolation or instruction to ignore embedded commands.
- Ingestion points: The
$ARGUMENTSvariable, which accepts a product description and a website link for the agent to process. - Boundary markers: There are no delimiters or explicit 'ignore embedded instructions' warnings around the interpolated user data.
- Capability inventory: The skill is granted
Bash(shell access),Write(file modification), andWebSearch(network retrieval) capabilities, which could be exploited if malicious content is processed. - Sanitization: The prompt contains no logic for filtering, escaping, or validating the data retrieved from external sources before it is analyzed.
Audit Metadata