The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill manages search indexing and filtering for product data, creating a surface for indirect prompt injection where data in the catalog could influence agent behavior.
Ingestion points: indexing.md describes the indexing process where product attributes, values, and localized content are fetched from the database or importer and stored in Elasticsearch.
Boundary markers: The skill does not define specific delimiters or instructions to the agent to treat search results or document payloads as untrusted content.
Capability inventory: The skill requires unopim-exec for Artisan command execution and uses the ElasticSearch facade for network-based queries and configuration updates.
Sanitization: The instructions include type coercion (e.g., filter_var for booleans) and normalization (sanitizeProductArray, ProductNormalizer) to ensure data matches expected mappings.
[COMMAND_EXECUTION]: The skill provides instructions for executing management commands that interact with the search infrastructure.
Evidence: indexing.md details the use of unopim:product:index, unopim:category:index, and unopim:elastic:clear. These commands are wrapped in unopim-exec and are used for routine maintenance tasks such as reindexing or clearing search indices.

unopim-search