The Agent Skills Directory

[METADATA_POISONING]: The skill metadata contains deceptive values, specifically listing a tested WordPress version of '7.0.2' and a last-updated date of '2026-08-06'. These entries are misleading as they reference non-existent versions and future dates.
[INDIRECT_PROMPT_INJECTION]: The skill establishes a process for importing large volumes of untrusted commerce data, creating an attack surface for indirect prompt injection if source data is malicious. * Ingestion points: Data adapters in fluent-cart-migrator/Classes/EDD3/ and fluent-cart/app/Modules/WooCommerceMigrator/. * Boundary markers: Instructions mandate isolated staging environments and manual reconciliation reports. * Capability inventory: The skill utilizes wp-cli for high-volume database writes and MigratorService::wipeMigratedData() for destructive resets. * Sanitization: Documentation describes monetary transformation but lacks details on sanitizing string fields against prompt injection.
[COMMAND_EXECUTION]: The skill encourages the use of wp-cli commands for data management and notes a specific safety bypass in version 1.0.0 where the --reset flag ignores the FLUENT_CART_DEV_MODE guard, increasing the risk of accidental production data loss.

fluentcart-migration