The Agent Skills Directory

[DATA_EXFILTRATION]: The skill instructions direct the agent to fetch and process content from external URLs provided by the user. While essential for performing a conversion audit, interacting with arbitrary external domains involves network operations to non-whitelisted sources.
[PROMPT_INJECTION]: The skill has an attack surface for indirect prompt injection as it ingests and processes untrusted data from external websites.
Ingestion points: The 'Fetch & Parse' step in SKILL.md extracts full rendered page content, including text and headings, from external URLs.
Boundary markers: The instructions do not define delimiters or provide 'ignore embedded instructions' warnings to separate external data from the agent's core instructions.
Capability inventory: The skill's primary capability is text analysis and report generation; there are no instructions for file-system writes or shell command execution based on the processed data, which mitigates the risk.
Sanitization: No sanitization, filtering, or validation of the fetched content is mentioned before the data is interpolated into the analysis prompt.

conversion-audit