Capability | What the skill does |
|---|
Search the OAC catalog | Searches OAC catalog items by name, including AI Agents, Workbooks, Datasets, Subject Areas, and Folders. Workbooks are handled as the OAC project catalog type. |
Map agents to datasets | Identifies dependencies between agents, datasets, and dataset columns. |
Create an agent from a dataset | Resolves the dataset from the OAC catalog, inspects and samples it through Oracle Analytics MCP, proposes a business scope, builds indexing rules, creates the agent, and optionally uploads document knowledge. |
Create an agent from a workbook | Parses an OAC workbook link or path, extracts its datasource, selected columns, calculations, filters, visual roles, parameters, and usage patterns, then proposes a workbook-aware agent scope before creation. |
Handle Subject Area sources | Detects Subject Areas from direct requests or workbook datasources, inspects and samples them, helps refine the scope, and proposes Logical SQL for a Local Subject Area before agent creation. |
Inspect datasets before agent creation | Uses Oracle Analytics MCP to describe and sample the dataset so the proposed scope, instructions, examples, AI Descriptions, and indexing are grounded in live metadata and data. |
Test an existing agent | Sends an approved set of exact questions to an existing agent using the bundled testing script and browser authentication. It captures one primary result per question and reports Pass or Failed, recommendations for failed questions, and optional additional optimizations. Testing consumes AIDP Units. |
Edit agent column subsetting independently | Fetches the existing agent definition and adds, removes, or replaces the datasource column subset while preserving unrelated agent settings. It validates that selected columns exist before saving and verifies the update afterward. |
Edit agent filters independently | Adds, removes, or updates fixed-value and approved expression filters without rebuilding the agent. New filters are hidden by default, filter visibility is considered during response testing, and parameter-bound filters are preserved unchanged. |
Manage AI Descriptions independently | Adds, changes, or removes dataset-level and selective column-level AI Descriptions while preserving unrelated indexing metadata. Proposed wording is reviewed before saving, and the dataset is re-indexed and verified afterward. |
Manage dataset indexing independently | Updates shared dataset indexing without recreating the agent, including Index Types, synonyms, and approved AI Descriptions, followed by re-indexing and status verification. |
Edit supplemental instructions independently | Fetches the current agent definition, changes only the targeted instruction fields, saves the complete updated definition, and verifies the update afterward. |
Edit welcome messages independently | Updates first-message or welcome-question content without rebuilding indexing or re-uploading document knowledge. |
Edit metadata independently | Updates the agent purpose, description, history setting, and catalog description while preserving existing datasource and document references. |
Manage document knowledge independently | Uploads, replaces, updates, or deletes document knowledge while checking existing documents and metadata to prevent duplicates. |
Use provided document knowledge | Uploads a user-provided .txt or .pdf file under 5 MB and verifies that vectorization completes successfully. |
Bulk edit instructions across agents | Applies targeted Supplemental Instruction updates across multiple existing agents, preserving unrelated fields and verifying every update. |
Preserve existing indexing when requested | Detects existing dataset indexing and can create or update the agent while leaving the shared indexing unchanged. |
Override indexing when requested | Replaces existing indexing with reviewed, agent-specific indexing rules when the user explicitly chooses to override it or no usable indexing exists. |
Work across OAC instances | Uses configuration-based OAC URLs instead of hardcoded tenant information. Browser authentication is the default and reuses or establishes a signed-in OAC session; token-file authentication remains available as an explicit fallback. |
