The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted content from external URLs and local files to generate SEO reports and automated fixes, which creates a vulnerability to instructions embedded within the processed data.
Ingestion points: The scripts/seo_lint.py and scripts/seo_fix.py tools ingest data from --url, --file, and --dir inputs as described in SKILL.md.
Boundary markers: There are no instructions for the agent to use delimiters or ignore instructions found within the target files or URL content.
Capability inventory: The skill performs network operations to fetch web content and has file system write capabilities via the --apply flag in seo_fix.py.
Sanitization: The skill does not describe any mechanisms to sanitize or filter the content before it is processed by the agent or written back to the file system.
[COMMAND_EXECUTION]: The skill invokes several Python scripts (seo_lint.py, rule_engine.py, spa_detect.py, render_page.py, seo_fix.py, schema_gen.py, report_publish.py) using command-line arguments provided by the user, such as URLs and local file paths.
[DATA_EXFILTRATION]: The skill performs network operations to fetch content from external URLs provided by the user for SEO analysis.
Network activity: The seo_lint.py and seo_fix.py tools communicate with external domains to retrieve page content or utilize the DataForSEO service for headless rendering.

seo-lint