aem-component

You receive a component name and produce a complete developer reference — every source file, AEM page, and dialog field for that component.

Performance target: ~10-15 tool calls total. No unnecessary agents or MCP calls.

1. Parse Input & Normalize

The argument is a component name. Accept any format and generate all search variations.

Step 1a: Extract base name

Strip resource type prefixes:

• project/components/content/hero → hero
• myapp/base/components/myapp-banner → myapp-banner

Read .ai/config.yaml for aem.component-prefix to know the project's naming prefix.

Step 1b: Generate search variations (enhanced with project.yaml)

From the base name, generate a regex pattern for grep that covers all naming conventions:

Given input {name}:

1. As-is: {name}
2. Remove hyphens/spaces: my-component → mycomponent
3. Add hyphens at word boundaries: mycomponent → my-component
4. Project prefix: <prefix>-{name} (from aem.component-prefix)
5. Strip prefixes: prefix-hero → also search hero

If .ai/project/project.yaml exists: Read platforms[].component-prefixes for additional prefix variations. For each prefix, generate {prefix}{name} as an additional variation. Example: prefixes [mycomp-, mycomp-bff-, canada-, corp-] → also search mycomp-{name}, mycomp-bff-{name}, canada-{name}, corp-{name}.

Build a single grep regex covering all variations.

If no argument provided, ask the user for the component name.

2. Search Local Docs (DIRECT — no agent)

Do NOT spawn an agent for this. Use direct Grep/Read calls — it's just 2-4 files.

Run these Grep calls in parallel (all variations in one regex pattern):

1. Grep .ai/project/component-index.md for all variations — the local repo component index (generated by /aem-init)
2. Grep .ai/project/component-index-project.md for all variations — the cross-repo enriched component catalog (if it exists)
3. Read .ai/config.yaml for project structure details — repos, component paths, SCM URLs