The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data fetched from external websites, creating a potential vector for indirect prompt injection attacks.
Ingestion points: scripts/fetch_page.py fetches HTML content from user-specified URLs.
Boundary markers: SKILL.md and references/skill-contract.md contain explicit directives to treat scraped content as evidence rather than instructions and to ignore any instructions found within external data.
Capability inventory: The skill can perform network requests (scripts/fetch_page.py) and write files to the local system (scripts/generate_markdown_report.py, scripts/generate_pdf_report.py, scripts/schema_generator.py).
Sanitization: scripts/fetch_page.py uses HTML unescaping for text extraction, and scripts/schema_generator.py employs standard JSON escaping when interpolating data into templates.
[EXTERNAL_DOWNLOADS]: The script scripts/generate_pdf_report.py references the reportlab library.
Evidence: The script includes a try-except block that suggests the user manually install the reportlab package using pip if PDF support is desired.
This is a well-known library for document generation, and the installation is not performed automatically by the script.
[DATA_EXFILTRATION]: The skill has the capability to perform network operations to arbitrary domains, which is necessary for its stated SEO analysis purpose.
The scripts/fetch_page.py utility implements security best practices to mitigate Server-Side Request Forgery (SSRF), including IP address validation to block private, loopback, and reserved ranges.
It also enforces resource limits such as maximum byte counts and timeouts for network requests.

content-gap-analysis