technical-blog-writing
Pass
Audited by Gen Agent Trust Hub on Apr 16, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill uses the
infshCLI via Bash to run applications for web searching (exa/search), image creation (infsh/html-to-image), and social media posting (x/post-create).- [EXTERNAL_DOWNLOADS]: Fetches installation guidelines from theinference-shGitHub repository, which is a trusted source.- [REMOTE_CODE_EXECUTION]: Utilizes theinfsh/python-executorto run a Python script that generates a benchmark chart. The script is statically defined in the skill and uses standard visualization libraries.- [PROMPT_INJECTION]: The skill exhibits a surface for indirect prompt injection by ingesting data from the internet via theexa/searchtool to assist in content research. - Ingestion points: External web content retrieved by the
exa/searchtool inSKILL.md. - Boundary markers: No specific delimiters or instructions to ignore embedded commands within the searched content are present.
- Capability inventory: Access to
Bashand the ability to run multipleinfshapplications (search, image generation, code execution, social posting). - Sanitization: No explicit sanitization or filtering of the external search output is implemented.
Audit Metadata