publish
Pass
Audited by Gen Agent Trust Hub on Oct 8, 2026
Risk Level: SAFECOMMAND_EXECUTIONDATA_EXFILTRATIONINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
- [COMMAND_EXECUTION]: The skill uses the Bash tool to perform network operations with curl and file manipulation with jq and zip.
- [DATA_EXFILTRATION]: User-selected content is uploaded to the external domain https://brewpage.app. This behavior is the primary intended function of the skill and is clearly documented, including warnings against publishing sensitive files like .env.
- [INDIRECT_PROMPT_INJECTION]: The skill interprets a free-form prompt to resolve file paths or text to be published.
- Ingestion points: User input provided in $ARGUMENTS as defined in SKILL.md.
- Boundary markers: Uses AskUserQuestion to confirm deployment parameters (namespace and password) before execution.
- Capability inventory: Utilizes Bash, Read, and Glob tools.
- Sanitization: Bash script templates use quoted heredocs (<<'BREWPAGE_EOF') to ensure that content to be published is treated as literal text and not executed as shell commands.
- [DYNAMIC_EXECUTION]: The skill constructs shell commands at runtime by interpolating resolved variables (content, namespace, TTL) into pre-defined bash templates.
Audit Metadata