The Agent Skills Directory

[PROMPT_INJECTION]: The skill processes audit data originating from external API sources (DataForSEO) stored in JSON format to generate PDF reports. This creates an indirect prompt injection surface where malicious data within the SEO audit could potentially target the agent or the PDF generation process.
Ingestion points: The skill reads audit data from ~/.claude/skills/seo/output/<domain>-audit.json.
Boundary markers: No explicit delimiters or boundary markers are defined to isolate external data from the generator's instructions.
Capability inventory: The skill uses Bash to execute the generator script and Write to create the PDF file.
Sanitization: The instructions do not mention sanitization or validation of the JSON content before processing.
[COMMAND_EXECUTION]: The skill executes local shell and Python scripts (preflight.sh, keyword_research.py, generate_pdf_report.py) using the Bash tool. It also uses chmod to manage file permissions and open to display the final report. These commands are scoped to the skill's internal directory and are consistent with its documented purpose.
[CREDENTIALS_UNSAFE]: The skill collects DataForSEO API credentials from the user and stores them in a local .env file (~/.claude/skills/seo/.env) with restricted filesystem permissions (chmod 600). This is categorized as safe practice for secret management in a local environment and does not pose a security risk.

seo-report-pdf