The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill possesses an attack surface for indirect prompt injection as it ingests and processes untrusted data from external sources.
Ingestion points: The skill reads WordPress post_content from pages and posts via respira_extract_builder_content and respira_read_page/respira_read_post to parse Divi shortcodes.
Boundary markers: There are no specific instructions for the agent to use delimiters or ignore instructions embedded within the shortcode content during parsing and conversion.
Capability inventory: The skill has extensive write capabilities, including respira_create_page_duplicate, respira_update_page, and respira_update_element, which could be abused if malicious instructions are successfully injected.
Sanitization: The instructions do not specify sanitization or validation of the extracted shortcode content before it is processed or written back to the site.
[DATA_EXFILTRATION]: The skill transmits usage telemetry to the developer's infrastructure.
Evidence: The Telemetry section specifies sending a POST request to https://www.respira.press/api/skills/track-usage containing skill slug, site context, and migration statistics upon completion.

migrate-divi-to-gutenberg