Searches for and retrieves existing visual media (images, logos, icons, photos, graphics, banners, thumbnails, hero images, backgrounds) from sources such as Salesforce CMS, Data 360 or any other source. Use this skill ANY TIME a user request involves finding, searching, getting, fetching, retrieving, grab, looking up, locating media. NEVER call search_media_cms_channels, search_electronic_media tools directly — always go through this skill first. This skill must be activated before any tool is
experience-content-media-search is a routing skill from Salesforce's official sf-skills repository that governs how an agent searches for and retrieves existing visual media (images, logos, icons, photos, banners, hero images, backgrounds) from sources such as Salesforce CMS and Data 360/Data Cloud. It solves the problem of an agent prematurely or blindly calling media search tools: the skill enforces that it activates first for any media search request and that a source is chosen before any tool runs.
The core workflow is deliberately constrained. On activation, the agent must first respond with a text-only numbered list of available search sources, built only from the tools actually present in its context (search_media_cms_channels for keyword search, search_electronic_media for Data 360 hybrid search, plus an "Other" option for a user-supplied URL or path). It must not probe or read MCP descriptors to discover tools, and must wait for the user to pick before executing the selected search. For keyword search it describes extracting up to ten concrete-noun keywords and descriptive taxonomies from the query, then presents results for selection and applies the chosen image to code.
The intended users are developers and builders working inside the Salesforce/Agentforce ecosystem who need to locate existing brand and content assets for their apps. It is explicitly scoped to finding existing media, not generating new images with AI, creating designs from scratch, or editing images, and it defers brand search to other tooling.
It routes media search and retrieval requests, presenting available search sources (Salesforce CMS keyword search, Data 360 hybrid search, or a user-supplied URL) and executing only the source the user selects.
It requires the search_media_cms_channels and/or search_electronic_media MCP tools. It shows only the options whose tools are actually present, and reports if none are configured.
No. It is strictly for finding existing media. Generating new AI images, creating designs from scratch, and editing existing images are out of scope.
It is a routing skill that requires an explicit user source selection first; it will not call any search tool or probe for available tools until you pick from the numbered list.
It analyzes your request, extracts up to ten concrete nouns (with domain synonyms) that would appear in image metadata, and extracts taxonomies describing styles, moods, and categories.
Quick Setup:
.claude/skills/Repository
forcedotcom/sf-skills