seo-content-writer
Pass
Audited by Gen Agent Trust Hub on Apr 15, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes local TypeScript utility scripts via the
tsxrunner to perform deterministic tasks such as workspace context discovery, keyword clustering, and article QA. These operations are restricted to the local environment and intended for skill automation.\n- [EXTERNAL_DOWNLOADS]: Communicates with the official Semrush API (api.semrush.com) to retrieve keyword metrics, competitor domain reports, and traffic analytics. These network operations are core to the skill's research functionality.\n- [DATA_EXFILTRATION]: Reads documentation and marketing materials from the local workspace to ground the AI in product-specific context. Snippets of this data are saved to theseo-content-output/directory for user review and auditing, rather than being transmitted to unauthorized remote servers.\n- [PROMPT_INJECTION]: The skill is subject to potential indirect prompt injection as it processes data from external APIs and workspace files to generate content briefs and draft instructions.\n - Ingestion points: Local file discovery in
scripts/workspace-context.tsand keyword/competitor data fromscripts/semrush-client.ts.\n - Boundary markers: Not implemented in the provided markdown templates (e.g.,
templates/brief-template.md).\n - Capability inventory: Shell command execution via
tsx, local file system writes, and network access to Semrush APIs.\n - Sanitization: Content is structured into markdown lists and tables but lacks explicit sanitization for adversarial prompt instructions.
Audit Metadata