marketing-pipeline-ai-content
Fail
Audited by Gen Agent Trust Hub on Oct 8, 2026
Risk Level: HIGHEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill instructions require cloning a repository from an untrusted individual GitHub account (
github.com/pennydinh/marketing-pineline-share.git). This source is not associated with the stated author or any known trusted organizations. - [REMOTE_CODE_EXECUTION]: The installation guide instructs the agent to run
npm installandnpm run devwithin the cloned repository. Executing scripts and installing dependencies from an unverified repository allows for arbitrary code execution on the user's machine. - [INDIRECT_PROMPT_INJECTION]: The skill possesses a vulnerability surface for indirect prompt injection.
- Ingestion points: Research data is fetched from external sources including TechCrunch, Twitter/X, and LinkedIn via the
crawlResearchmodule. - Boundary markers: The prompt template (
Based on this research: ${JSON.stringify(config.research)}) does not use delimiters, tags, or system instructions to isolate the external data. - Capability inventory: The system makes network calls to AI providers (Anthropic, OpenAI) and performs local file system operations via the Remotion rendering engine.
- Sanitization: There is no evidence of sanitization or filtering of the content retrieved from external websites before it is interpolated into the AI generation prompt.
Recommendations
- AI detected serious security threats
Audit Metadata