The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it retrieves and processes content from untrusted external domains.
Ingestion points: The skill performs webfetch operations on /robots.txt, /sitemap.xml, and various sample pages from target domains.
Boundary markers: There are no specified delimiters or instructions to ignore potential commands embedded within the fetched web content.
Capability inventory: The agent is instructed to write files to disk (robots.txt) and execute a reporting script (scripts/report_publish.py).
Sanitization: The instructions do not define any sanitization, filtering, or validation steps for the data retrieved from external URLs before it is processed by the agent.
[COMMAND_EXECUTION]: The skill invokes a local Python script as part of its normal operation for generating client reports.
Evidence: The skill body specifies running python scripts/report_publish.py <report>.md to produce branded HTML and PDF outputs.

robots-advisor