# DeepRAG in a Low-Code Agent — Planning

## When to Use

Pick this when:

- Project has `agent.json` with `"type": "lowCode"` (standalone) **or** the agent is inline inside a Maestro Flow (`uipath.agent.autonomous` node)
- User is building in Studio Web Agent Builder, no Python
- Agent receives runtime attachments (files uploaded in chat) and must research / synthesize across them
- Expected output is a single grounded narrative answer

Confirm DeepRAG is the right mode first — see [../../../../context-grounding-patterns.md](../../../../context-grounding-patterns.md).

For coded agents (Python, LangGraph) → [../../../../coded/capabilities/deeprag/planning.md](../../../../coded/capabilities/deeprag/planning.md).

## Inputs You Need Before Building

| Input | Why | Source |
|---|---|---|
| Agent project shape | Standalone vs. inline-in-flow — affects where `resource.json` lives | Inspect `agent.json` and the parent solution |
| Attachment ingress | The `deep-rag` tool consumes runtime-uploaded attachments — confirm the agent has an attachment input wired | Studio Web schema / `entry-points.json` |
| Prompt wiring | The agent's system prompt or per-call prompt must mention when to invoke the tool | Author / Studio Web |
| Other built-in tools | Do not enable `deep-rag` AND `analyze-attachments` if a single research pass suffices — pick one | User intent |

## Tool Resource Shape

Built-in tools are declared in `resources/<name>/resource.json` with `$resourceType: "tool"`, `type: "internal"`, `referenceKey: null`, and `properties.toolType: "deep-rag"`. See [impl-json.md](impl-json.md) for the exact JSON.

Built-in tools accept these `toolType` values: `analyze-attachments`, `load-attachments`, `deep-rag`, `batch-transform`. Anything else fails `uip agent validate`.

### Tool resource vs context-index resource

Two valid shapes for enabling DeepRAG on a low-code agent. This skill documents the **built-in tool** shape (the agent invokes DeepRAG through its tool-calling loop on runtime attachments). The alternative is a **context-index resource** that wires DeepRAG as a retrieval mode on a pre-built ECS index — `$resourceType: "context"`, `contextType: "index"`, `retrievalMode: "deeprag"` (lowercase, no hyphen — `uip agent validate` accepts camelCase but Studio Web silently drops the resource on import), with `citationMode` (`"Inline"` or `"Skip"`) set on the resource. Use the context-index form when the documents live in a stable, pre-built index reused across runs and the agent should query it transparently as context; use the tool form (this skill) when the documents are runtime attachments uploaded per conversation.

## Critical Decisions

| Decision | Rule |
|---|---|
| `deep-rag` vs `analyze-attachments` | Prefer `deep-rag`. It handles larger files, returns citations (with optional bounding-box anchors), and produces denser output. `analyze-attachments` is one-shot synthesis with lower page limits — reach for it only when the document is small and the task is a single-shot extraction. |
| `deep-rag` vs `load-attachments` | `load-attachments` only makes attachment text available to the agent; `deep-rag` runs an iterative synthesis pass. Use `load-attachments` when the agent will reason directly over short contents; use `deep-rag` for long / multiple docs. |
| Standalone agent vs inline-in-flow | Same `resource.json` shape for both. The flow wiring differs — inline requires an edge from the agent's `tool` port to the tool node's `input` port. See [impl-json.md](impl-json.md). |

## Bindings / Permissions

DeepRAG runs in the folder context of the running agent. The tool requires the index permission in that folder — same rule as the coded surface. If the agent is published to a folder where the user lacks rights, DeepRAG calls fail with 403. Use the personal workspace (or a folder the agent's runtime identity has rights in) for self-serve.

## Hand-off

Once planning is complete, implement per [impl-json.md](impl-json.md).