Oracle Analytics

View saved version →

---
name: oac-analytics-qa
description: Answer business and analytical questions with Oracle Analytics Cloud using the OAC MCP tools. Use when a user asks for totals, trends, comparisons, breakdowns, KPIs, filtering, aggregation, or other data questions that require discovering an OAC data source, inspecting metadata, executing Oracle Analytics Logical SQL, and explaining the returned results. Keep this workflow read-only and use Logical SQL as the default query path when the analytical context requires it.
---

# Oracle Analytics Cloud Analytics Q&A

## Configured instance

Use only the Oracle Analytics MCP connection attached to the selected configured instance. Ground requests in that instance's live tools, metadata, identifiers, and permissions. Do not fall back to the template example URL, a remembered demo account or catalog path, or another instance's connection. If no instance is configured, explain that a workspace administrator must configure one before this workflow can run.

Use the Oracle Analytics Cloud MCP server as the execution layer. Use the runtime-exposed tool schema as the source of truth for arguments, required fields, and return values. Raw MCP tool names are listed below; the host may normalize their identifiers.

## Workflow

1. Interpret the analytical question.
   - Preserve the user's requested measures, dimensions, filters, date range, comparison period, and level of detail.
   - Do not add business definitions or filters that the user did not request or that OAC metadata does not establish.

2. Resolve the data source.
   - If the relevant subject area or dataset is already known from current conversation context, reuse it.
   - Otherwise call `oracle_analytics-find_matching_datasources` with the user's analytical intent.
   - If multiple candidates are returned, use metadata to distinguish them when possible. Ask the user only when the available tools cannot resolve a material ambiguity.

3. Ground the query in metadata.
   - Call `oracle_analytics-describe_data` before composing a query when table or column structure is not already established by current OAC tool output.
   - Use only returned table, column, subject-area, dataset, and related metadata. Never fabricate names, relationships, data types, or business meanings.

4. Choose the query tool.
   - Use `oracle_analytics-execute_logical_sql` for standard analytical queries or when the user explicitly requests Oracle Analytics Logical SQL or when the requested analysis depends on Logical SQL semantics established by the available context.
   - Do not silently substitute one query language for the other when the distinction matters.
   
5. Execute the query.
   - Match the aggregation and grain to the user's question.
   - Avoid unnecessary detail or unbounded row retrieval when an aggregate answer satisfies the request.
   - If the query fails, revise it only from returned metadata or tool error information. Do not guess at schema changes.

6. Present the answer.
   - State the result clearly and concisely.
   - Identify material filters, date ranges, grouping, or assumptions that came from the request or tool output.
   - Distinguish returned data from interpretation. Do not present an interpretation as a field returned by OAC.
   - Report units and scale only when source metadata or an authoritative data dictionary explicitly establishes them. A column name containing "Rate" or "Ratio" does not establish a percentage. If units or scale are unknown, preserve the returned numeric value and state the limitation; do not append `%`, rescale the value, or use inferred units in charts or follow-ups.
   - If the data is incomplete or ambiguous, explain the limitation rather than filling gaps.

## MCP tools

- `oracle_analytics-find_matching_datasources`: Match a natural-language analytical question to relevant OAC data sources.
- `oracle_analytics-describe_data`: Retrieve table and column metadata for subject areas or datasets.
- `oracle_analytics-execute_oac_ansi_sql`: Execute OAC ANSI SQL when explicitly requested and supported by the live tool schema; otherwise use Logical SQL as the default.
- `oracle_analytics-execute_logical_sql`: Execute Oracle Analytics Logical SQL and return query results.

## Boundaries

- Keep this skill read-only. Do not invoke catalog mutation or permission-management tools as part of an analytical Q&A request.
- Do not claim a query succeeded until the MCP tool returns a successful result.
- Do not invent MCP parameters. Follow the live runtime schema.
- The Oracle Analytics Cloud MCP integration is in Preview; avoid promising behavior beyond the currently exposed tools.

## Example requests

- "Show revenue by region for Q2."
- "Which product categories grew the most year over year?"
- "Use Logical SQL to return monthly bookings by sales organization."
- "What is the average order value for enterprise customers this quarter?"

View saved version →

---
name: oac-catalog-management
description: Manage Oracle Analytics Cloud catalog content through the OAC MCP tools. Use when a user wants to create or replace supported workbook or connection content, export a workbook to PNG or PDF, create a catalog folder, copy an item, move an item, or delete an item. Use catalog search to resolve targets when needed. Treat create, replace, copy, move, and delete operations as production-impacting mutations that require explicit approval immediately before execution; export is non-mutating but starts an export job.
---

# Oracle Analytics Cloud Catalog Management

## Configured instance

Use only the Oracle Analytics MCP connection attached to the selected configured instance. Ground requests in that instance's live tools, metadata, identifiers, and permissions. Do not fall back to the template example URL, a remembered demo account or catalog path, or another instance's connection. If no instance is configured, explain that a workspace administrator must configure one before this workflow can run.

Use this skill for catalog organization and content lifecycle actions. Use the runtime-exposed MCP schema as the source of truth for arguments, identifiers, supported options, and result fields.

## Preflight

1. Identify the exact target object and requested operation.
   - Check whether the requested outcome is supported before starting any export or mutation. Email delivery and recipient-directory lookup are outside this plugin's tools.
   - For a request to email an export, explain this limitation and ask whether the user wants a supported PNG or PDF export. Do not start an export job or select another plugin before that separate choice.
2. If the target or destination is not already established by current tool output or an unambiguous identifier, use `oracle_analytics-search_catalog` to locate it.
3. For copy or move, establish both source and destination.
4. For create or replace, establish the supported content type and target catalog location.
5. For export, establish the workbook and requested format. The documented export formats are PNG and PDF.
6. Never infer overwrite, replacement, recovery, retention, or destination behavior that is not established by the live tool schema.

## Approval rule for mutations

Before calling any modifying tool, present the exact operation as understood and require the product's explicit approval flow. Modification tools require `userApproved: true` or an explicit approval flow.

Treat these as approval-gated production-impacting actions:

- `oracle_analytics-save_catalog_content`
- `oracle_analytics-create_catalog_folder`
- `oracle_analytics-copy_catalog_item`
- `oracle_analytics-move_catalog_item`
- `oracle_analytics-delete_catalog_item`

Do not set or imply approval until the user has approved the exact operation through the supported flow. A broad request to "manage" or "clean up" content is not approval for an unspecified mutation.

## Operation workflows

### Create or replace content

Use `oracle_analytics-save_catalog_content` only for supported workbook or connection catalog content.

- State whether the action will create or replace content when that distinction is known.
- Do not silently overwrite an existing item.
- Require explicit approval immediately before execution.

### Export a workbook

Use `oracle_analytics-export_workbook` to export a workbook to PNG or PDF.

- Export does not modify the catalog, but it starts an export job.
- No catalog-mutation approval is required by the documented tool table.
- Report export or job status only from the tool response.

### Create a folder

Use `oracle_analytics-create_catalog_folder`.

- Confirm the folder name and target location.
- Require explicit approval immediately before execution.

### Copy an item

Use `oracle_analytics-copy_catalog_item`.

- Confirm the source item and destination.
- Require explicit approval immediately before execution.

### Move an item

Use `oracle_analytics-move_catalog_item`.

- Confirm the source item and destination.
- Require explicit approval immediately before execution.

### Delete an item

Use `oracle_analytics-delete_catalog_item`.

- Identify the exact item and location.
- Do not infer that an item is safe to delete because it appears unused.
- Do not invent soft-delete, recycle-bin, retention, or recovery behavior.
- Require explicit approval immediately before execution.

## MCP tools

- `oracle_analytics-search_catalog`: Locate catalog objects before an operation when needed. Experimental.
- `oracle_analytics-save_catalog_content`: Create or replace workbook or connection catalog content. Experimental; modifies OAC; approval required.
- `oracle_analytics-export_workbook`: Export a workbook to PNG or PDF. Experimental; starts an export job; does not modify the catalog.
- `oracle_analytics-create_catalog_folder`: Create a catalog folder. Experimental; modifies OAC; approval required.
- `oracle_analytics-copy_catalog_item`: Copy a catalog item. Experimental; modifies OAC; approval required.
- `oracle_analytics-move_catalog_item`: Move a catalog item. Experimental; modifies OAC; approval required.
- `oracle_analytics-delete_catalog_item`: Delete a catalog item. Experimental; modifies OAC; approval required.

## Result handling

- Do not claim success until the MCP tool confirms it.
- Report the resulting object, destination, export status, or error using only returned information.
- If an export failed or no artifact is available, do not draft text claiming the file is attached or ready for delivery.
- Preserve uncertainty and tool errors rather than masking them.

## Boundaries

- Do not change permissions in this skill; use the governance and permissions workflow for ACL changes.
- Do not invent MCP parameters or unsupported catalog object types.
- Do not provision users, update source-database records, send email, or claim that these actions were completed.
- The OAC MCP integration is in Preview and the catalog-management tools listed above are experimental.

## Example requests

- "Create a folder named FY27 Planning under Shared Folders."
- "Move the Pipeline workbook into Sales Operations."
- "Export this workbook as PDF."
- "Delete the old forecast workbook."

View saved version →

---
name: oac-content-discovery
description: Discover and inspect Oracle Analytics Cloud content and data sources using the OAC MCP tools. Use when a user wants to find workbooks, folders, datasets, connections, subject areas, analyses, reports, dashboards, or scripts; determine which OAC data source can answer a natural-language question; or inspect tables and columns for a known dataset or subject area. Keep discovery read-only, return only tool-supported metadata, and use search, source matching, and metadata inspection in the smallest sequence needed to answer the request.
---

# Oracle Analytics Cloud Content Discovery

## Configured instance

Use only the Oracle Analytics MCP connection attached to the selected configured instance. Ground requests in that instance's live tools, metadata, identifiers, and permissions. Do not fall back to the template example URL, a remembered demo account or catalog path, or another instance's connection. If no instance is configured, explain that a workspace administrator must configure one before this workflow can run.

Use this skill to help users locate OAC content and understand available analytical data before querying or modifying anything. Use the runtime MCP schema as the source of truth for arguments and response fields.

## Choose the discovery path

1. For a named or described catalog object, call `oracle_analytics-search_catalog`.
   - Search the user's stated criteria without inventing catalog paths or object names.
   - The documented searchable object types include workbooks, folders, datasets, connections, subject areas, analyses, reports, dashboards, and scripts.
   - Return enough tool-provided context to distinguish similar matches.

2. For a business question where the user does not know the source, call `oracle_analytics-find_matching_datasources`.
   - Preserve the analytical intent of the user's question.
   - Explain source relevance only from information returned by the tool.
   - Do not infer unavailable measures, dimensions, relationships, or permissions.

3. For a known subject area or dataset, call `oracle_analytics-describe_data`.
   - Retrieve table and column metadata needed to understand its structure.
   - Do not fabricate columns, data types, joins, relationships, or business definitions.

4. Chain tools only when useful.
   - Search -> describe when the user first needs an object located and then wants its structure.
   - Match source -> describe when a natural-language question needs a suitable source and then schema grounding.
   - Stop once the user's discovery question is answered; do not execute SQL unless the request also asks for analysis.

## MCP tools

- `oracle_analytics-search_catalog`: Search catalog objects. This tool is experimental.
- `oracle_analytics-find_matching_datasources`: Match a natural-language question to relevant OAC data sources.
- `oracle_analytics-describe_data`: Retrieve table and column metadata for subject areas or datasets.

## Result handling

- Prefer a short ranked-by-relevance presentation only when the tool itself provides relevance information; otherwise present matches without inventing a ranking.
- Clearly distinguish exact matches from partial or ambiguous matches when the returned metadata supports that distinction.
- If no match is returned, say that no matching item was found under the supplied criteria. Do not manufacture a fallback object.
- Do not expose metadata that the MCP tool did not return.

## Boundaries

- Treat all discovery operations as read-only.
- Do not invoke catalog mutation, export, query execution, or ACL update tools unless the user's request extends beyond discovery and another applicable workflow governs the action.
- Do not invent MCP parameters. Use the host's live tool schema.
- `oracle_analytics-search_catalog` is experimental in the Oracle documentation, and the OAC MCP integration is in Preview.

## Example requests

- "Find the workforce planning workbook."
- "Which dataset should I use to answer a question about renewal revenue?"
- "What columns are available in the Sales subject area?"
- "Show me dashboards related to pipeline."

View saved version →

---
name: oac-governance-permissions
description: Manage Oracle Analytics Cloud catalog permissions through the OAC MCP tools. Use when a user asks to grant, revoke, or otherwise change access to a catalog item, or when a permission-management workflow needs to locate the exact target item before updating its ACL. Treat ACL updates as security-sensitive, production-impacting catalog mutations. Require explicit approval immediately before execution, never infer users, groups, roles, inheritance, or permission semantics, and do not claim that current permissions were inspected unless a connected tool actually returned them.
---

# Oracle Analytics Cloud Governance and Permissions

## Configured instance

Use only the Oracle Analytics MCP connection attached to the selected configured instance. Ground requests in that instance's live tools, metadata, identifiers, and permissions. Do not fall back to the template example URL, a remembered demo account or catalog path, or another instance's connection. If no instance is configured, explain that a workspace administrator must configure one before this workflow can run.

Use this skill for catalog access-control changes. Use the runtime-exposed MCP schema as the authoritative source for supported ACL arguments, principal identifiers, permission values, and response fields.

## Workflow

1. Identify the exact catalog target.
   - If the target is not already established by an unambiguous identifier or current tool output, use `oracle_analytics-search_catalog` to locate it.
   - Do not infer a target from a similar name when multiple matches exist.

2. Establish the requested permission change.
   - Preserve the principal, permission, scope, and intended action exactly as established by the user and supported tool schema.
   - Do not invent or infer user IDs, group IDs, application roles, inheritance behavior, permission names, or ACL merge/replace semantics.

3. State the exact intended change.
   - Make clear which catalog item will be changed and what access change will be attempted.
   - If the live tool semantics indicate a replace-versus-merge distinction, surface that distinction before approval.

4. Require explicit approval.
   - `oracle_analytics-update_catalog_acl` modifies Oracle Analytics Cloud and requires `userApproved: true` or an explicit approval flow.
   - Obtain approval immediately before invoking the ACL update.
   - Do not interpret a general request to review or manage permissions as approval for a specific ACL mutation.

5. Execute the update.
   - Invoke `oracle_analytics-update_catalog_acl` only with values supported by the live schema and grounded in user intent or prior tool output.
   - Do not broaden access beyond what the user requested.

6. Report the outcome.
   - Do not claim success until the tool confirms it.
   - Report resulting permissions or status only when returned by the tool.
   - Preserve errors or uncertainty rather than filling gaps.

## MCP tools

- `oracle_analytics-search_catalog`: Locate the target catalog item when necessary. Experimental and read-only.
- `oracle_analytics-update_catalog_acl`: Update catalog permissions. Experimental; modifies OAC; approval required.

## Important limitation

The documented OAC MCP tool set does not list a dedicated read-only ACL inspection tool. Do not claim that this skill can enumerate or audit existing permissions unless the active runtime exposes additional supported functionality or the ACL update tool returns that information. If a user asks only "who has access?", state the limitation instead of fabricating a permission view.

## Security rules

- Prefer the smallest change that exactly matches the user's request.
- Never grant broader access as a convenience or workaround.
- Do not make assumptions about inherited permissions or effective access.
- Do not convert names to identities unless a connected tool explicitly resolves them.
- Do not invent MCP parameters. Follow the live runtime schema.
- The OAC MCP integration is in Preview, and both catalog search and ACL update are experimental tools.

## Example requests

- "Give the Finance Analysts group access to this workbook."
- "Remove access for this role from the Planning folder."
- "Change the permissions on the Q4 forecast workbook."