> ## Documentation Index
> Fetch the complete documentation index at: https://mantle.rasa.com/llms.txt
> Use this file to discover all available pages before exploring further.

# What's Next

> Planned features and capabilities coming to Mantle.

These features are designed but not yet built. Some already appear in
`integrations.yml` as configuration that parses and validates without a runtime
consumer; those are called out below.

## Responses and questions

**Variant selection by channel.** Variants are already selected at runtime by
`condition` (evaluated against memory); selecting a variant by the active
channel is planned.

**`ask_<memory_entry>` hooks** for dynamically generated questions, so a skill
can compute the wording of a `collect:` prompt at runtime.

## Integrations

**Knowledge stores.** `knowledge_stores:` (Qdrant, Milvus, external FAISS)
validates but is never queried. Today `search_knowledge` reads only the local
index built from `references/**/*.md`.

**Per-skill retrieval scoping.** A skill's own `references/` are already
indexed and tagged by scope alongside the shared index; searching a skill's own
references ahead of, or instead of, the shared index is planned.

## Remote tool hosting

**Dedicated tool server.** A tool server that the runtime connects to over MCP,
letting you host an agent's tools separately from the runtime, is planned. Today
Mantle connects out to remote MCP tool servers you declare under `mcp_servers:`
in `integrations.yml`.

**MCP extras.** `pre_call_hook` and `meta_map.from_slots` are not supported.
Use `meta_map.static` for fixed `_meta`, or `meta_map.from_memory` to send
skill or project memory at call time.

## Agent-to-agent

**A2A skill exposure** will let you expose a skill so other agents can invoke it
over the A2A protocol.


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.