web-quality-audit
Pass
Audited by Gen Agent Trust Hub on Mar 1, 2026
Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [PROMPT_INJECTION]: The skill metadata identifies 'addyosmani' as the author, whereas the provided author context is 'thoreinstein'. This inconsistency in authorship metadata can be misleading regarding the skill's origin.
- [COMMAND_EXECUTION]: The
scripts/analyze.shscript is vulnerable to argument injection. It passes the$filevariable directly togrepwithout using the--delimiter. An attacker could use filenames starting with hyphens to inject flags into thegrepcommand, potentially allowing the reading of unauthorized local files. - [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it processes untrusted HTML files.
- Ingestion points: The
scripts/analyze.shscript reads and analyzes the content of HTML files provided as input. - Boundary markers: While audit results are structured as JSON, the agent's subsequent processing of the content lacks explicit delimiters to ignore embedded instructions.
- Capability inventory: The skill utilizes shell commands (
grep,find) to inspect local files and report findings to the agent. - Sanitization: No sanitization is performed on the content read from the audited files before it is processed or presented to the agent.
Audit Metadata