Opal tools

  • Updated

This page lists the built-in Optimizely Opal tools provided and maintained by Optimizely. You and Opal can call these tools if they are enabled. Review the tools so you can see what actions are available for you and your workflows.

Click a tool's name to expand it and learn when to use it, its required and optional parameters, and example prompts for calling the tool. If you do not provide a required parameter, Opal prompts you for it.

Agents

ask_about_agent_execution – Answers a plain-language question about a single agent run and detects whether the run belongs to a workflow agent or a specialized agent.
  • When to use
    • Ask a question about a run instead of reading the full execution record.
    • Confirm whether a specific step in a run finished or failed.
    • Identify the slowest step in a workflow run.
    • Summarize what an agent produced during a run.
    • Investigate a run when the agent type is unknown.
  • Parameters
    • execution_id – The ID of the run to ask about. Opal accepts a workflow run or a specialized agent run.
    • question – The question about the run, written in plain language.
  • Example prompts
    • For execution exec-8821, did the email step finish?
    • Which step took the longest in run exec-4410?
    • Summarize what happened in execution exec-9302.
    • Tell me why execution exec-1157 failed.
ask_agent_input – Displays a form in chat that gathers the input a specialized agent needs before the agent runs.
  • When to use
    • Start an agent without knowing which details it requires.
    • Enter several values at one time instead of answering separate questions.
    • Attach a file or build a list that an agent accepts as input.
    • Review and adjust values that Opal fills in for you before the run starts.
  • Parameters
    • agent – The specialized agent that receives the input. Opal identifies the agent from the name or handle in your request.
    • parameters – The fields the form displays, including each field label, help text, and input type. Opal builds these fields from the agent definition.
  • Example prompts
    • Run the case study agent and ask me for what it needs.
    • I want to use the competitive insights agent. Show me a form.
    • Set up the creative brief agent, but let me fill in the details first.
create_or_edit_agent – Drafts a specialized agent or edits an existing one, then displays an approval card that saves the agent after you confirm it.
  • When to use
    • Build an agent for a task you repeat often.
    • Change the instructions, inputs, tools, or output format of an existing agent.
    • Rename an agent or shorten its handle.
    • Add quality criteria and a passing score to an agent.
    • Turn a single-run agent into a multi-turn conversation agent.
  • Parameters
    • (Optional) agent_id – The handle of the agent to edit. Omit this value to create an agent.
    • (Optional) name – The display name of the agent, such as *Support Ticket Triager*. Required when you create an agent.
    • (Optional) description – A short summary of what the agent does and when to use it.
    • (Optional) prompt_template – The instructions the agent follows on every run. Write the instructions in second person and reference inputs with the [[parameter_name]] format. Required when you create an agent that does not use conversation mode.
    • (Optional) parameters – The inputs the agent accepts. Omit this value for an agent that needs no input.
    • (Optional) output – The format the agent returns, such as a report or a structured response.
    • (Optional) enabled_tools – The tools the agent calls during a run. Each tool must exist in your Opal instance.
    • (Optional) inference_type – The model tier. Use simple for most tasks, complex or complex_with_thinking for multi-step reasoning, and code for code-heavy work. Defaults to simple.
    • (Optional) conversation_mode – Set to true for a multi-turn chat agent and false for a single-run task agent. Defaults to false.
    • (Optional) system_prompt_template – The system instructions for a conversation agent. Opal applies this value only when conversation_mode is true.
    • (Optional) initial_message – The greeting a conversation agent displays before you type.
    • (Optional) enable_auto_tool_prediction – Lets the agent select relevant tools during a run, in addition to the tools in enabled_tools. Defaults to false.
    • (Optional) enable_personal_context – Adds your saved profile and preferences to the agent instructions. Defaults to false.
    • (Optional) enable_auto_skill_prediction – Applies organization skills that match the agent instructions. Defaults to false.
    • (Optional) enable_auto_skill_prediction_by_tools – Applies organization skills that match the tools of the agent. Defaults to false.
    • (Optional) baseline_evaluation_score – The score from 0 through 100 that a run must reach to pass. Opal displays this value with the evaluation criteria on the Quality tab.
    • (Optional) evaluation_criteria – The quality standards Opal scores each run against, such as Response answers the question directly. Opal displays these standards on the Quality tab. Pass an empty list to clear every standard.
    • (Optional) input_guardrails – The trained guardrail version that screens input before a run.
    • (Optional) is_enabled_in_chat – Makes the agent available in Opal Chat. Defaults to true.
    • (Optional) llm_provider – The model provider, either google or anthropic. Omit this value to accept the default provider for the selected inference_type.
    • (Optional) cwd – The working directory the agent uses for files and personal context. The path must start with /workspace or /context. A /context path requires enable_personal_context set to true.
    • (Optional) file_urls – The reference files the agent uses as knowledge, such as templates or brand guidelines. Upload each file first. Pass an empty list to remove every attached file.
    • (Optional) new_agent_id – The replacement handle for an existing agent. Use letters, digits, hyphens, and underscores. To change only the display name, use name instead.
    • (Optional) visibility – The audience for an agent you create. Use private to restrict the agent to yourself, or shared to release it to everyone in your instance.
  • Example prompts
    • Create an agent that sorts support tickets by urgency and drafts a first reply.
    • Edit my release notes reviewer so it also flags passive voice.
    • Build a conversation agent that coaches product managers on user stories.
    • Add quality criteria to my changelog agent with a passing score of 80.
    • Rename the handle for my LinkedIn post writer to li-post-writer.
execute_specialized_agent – Runs a specialized agent with the input values you supply.
  • When to use
    • Run an agent when you already have every required value.
    • Start an agent you named directly in chat.
    • Send a file, a list, or another structured value to an agent.
    • Repeat a run with different values and compare the results.
  • Parameters
    • agent_id – The handle of the specialized agent to run.
    • parameters – The input values the agent accepts, supplied as structured data.
  • Example prompts
    • Run my competitive insights agent for Acme Corp.
    • Run the release notes reviewer on the notes I pasted.
    • Use the changelog agent on this list of API updates.
    • Run the alt text agent on https://www.optimizely.com/products.
execute_workflow_agent – Runs a workflow agent from a plain-language description of the work.
  • When to use
    • Start a workflow agent that coordinates several steps.
    • Launch a chained process, such as a campaign build or a release.
    • Start a workflow and capture the run ID for later review.
    • Run a process that calls several specialized agents in order.
  • Parameters
    • agent_id – The handle of the workflow agent to run.
    • input_text – The description of the work, written in plain language.
  • Example prompts
    • Run my changelog workflow on the two files I uploaded.
    • Run the landing page workflow for our enterprise segments.
    • Start the review workflow for the alert that arrived this morning.
    • Start the release automation workflow for the August release.
get_agent – Returns the full definition of one specialized agent.
  • When to use
    • Review the instructions, inputs, output, and tools of an agent.
    • Read the configuration of an agent before you change it.
    • Confirm which tools an agent calls.
    • Look up an agent from the ID in its Opal address.
    • Document the configuration of an existing agent.
  • Parameters
    • agent_id – The handle of the agent or the ID from its Opal address. Opal resolves either value.
  • Example prompts
    • Show me the full definition of my SME tracker agent.
    • Which tools does my doc review agent use?
    • Get the configuration for the agent I have open.
    • What inputs does my page builder agent accept?
get_agent_execution – Returns the details of one run and detects whether the run belongs to a workflow agent or a specialized agent.
  • When to use
    • Troubleshoot a run when you have the run ID but not the agent type.
    • Review the steps, trigger, and status of a workflow run.
    • Check the status, duration, token use, credits, and response of a run.
    • Find the error that ended a failed run.
    • Collect the memory IDs that reveal how an agent reached its result.
  • Parameters
    • execution_id – The ID of the run to troubleshoot. Opal accepts a workflow run or a specialized agent run.
    • (Optional) full – Returns every step of a workflow run with complete detail. Opal ignores this value for a specialized agent run. Defaults to false, which returns a short overview.
  • Example prompts
    • Troubleshoot execution exec-7741.
    • Show me the full detail for workflow run exec-2290.
    • What was the final response for run exec-6653?
    • How many credits did execution exec-3312 use?
get_workflow_agent_step_execution – Returns the full detail of one workflow step, including its input values, status, and any error.
  • When to use
    • Review the input values a workflow passed to a step.
    • Read the response a specialized agent returned inside a workflow step.
    • Check how a conditional step evaluated and which branch it followed.
    • Review the items and per-item results of a loop step.
    • Collect the memory IDs that reveal how a workflow handled one step.
  • Parameters
    • step_execution_id – The ID of the step to review. A workflow run record lists this ID for each step.
  • Example prompts
    • Show me the detail for step execution step-exec-4471.
    • Which values did the workflow pass to that step?
    • Which branch did the conditional step follow?
    • How many times did the loop step run?
list_agent_executions – Lists workflow runs and specialized agent runs together, starting with the most recent.
  • When to use
    • Review recent activity when you do not have a run ID.
    • Find every failed run within a date range.
    • Review the run history of one agent.
    • Confirm that a scheduled workflow ran as expected.
    • Include the runs that a workflow started, alongside the runs you started.
  • Parameters
    • (Optional) agent_id – Limits results to one agent handle. To search by name instead, use agent_name_contains.
    • (Optional) agent_name_contains – Limits results to agents whose name contains this text. Use this value when you remember only part of the name.
    • (Optional) agent_type – Limits results to workflow runs or specialized runs. Omit this value to list both.
    • (Optional) statuses – Limits results to one or more statuses. Workflow runs use pending, running, completed, failed, and cancelled. Specialized runs use running, waiting, completed, failed, and cancelled.
    • (Optional) after – Limits results to runs that started at or after this date and time.
    • (Optional) before – Limits results to runs that started at or before this date and time.
    • (Optional) include_workflow_triggered – Adds the runs that a workflow started, including nested workflows. Defaults to false, which lists only the runs you started directly.
    • (Optional) limit – The number of runs per page. Defaults to 25, with a maximum of 100.
    • (Optional) page_key – The marker that returns the next page of results. Omit this value for the first page.
  • Example prompts
    • Show me every failed run from today.
    • List the last 10 runs of my SME tracker agent.
    • Why does my most recent run keep failing?
    • List all workflow runs from the past week, including nested workflows.
list_workflow_agent_step_executions – Lists every run of a workflow step, starting with the oldest.
  • When to use
    • Review a step that ran more than one time in a single workflow run.
    • Narrow the steps of a workflow run to the ones that failed.
    • List every pass of a loop step.
    • Check how many times a step retried before it finished.
    • Trace each step run back to the loop pass that started it.
  • Parameters
    • workflow_execution_id – The ID of the workflow run that contains the steps.
    • (Optional) step_id – Limits results to one step in the workflow definition.
    • (Optional) statuses – Limits results to one or more of pending, running, completed, failed, cancelled, skipped, and waiting.
    • (Optional) after – Limits results to steps that started at or after this date and time.
    • (Optional) before – Limits results to steps that started at or before this date and time.
  • Example prompts
    • List every step in workflow run exec-5520.
    • Show me only the failed steps in that workflow run.
    • List every pass of the landing page step.
    • Did any step in this run retry?
lookup_approved_plans – Returns approved workflow plans that match the work a goal-driven workflow received, so the workflow reuses a proven approach.
  • When to use
    • Reuse an approved plan instead of planning a run from the start.
    • Adapt a proven plan to similar incoming work.
    • Check whether a goal-driven workflow has approved plans on record.
    • Produce steadier results across repeated runs of the same workflow.
  • Parameters
    • workflow_agent_id – The handle of the workflow agent that requests the plans. Opal searches only the approved plans of that agent.
    • trigger_envelope – The incoming work that Opal matches approved plans against.
    • (Optional) k – The number of plans to return, closest match first.
  • Example prompts
    • Check for approved plans before you build a new one.
    • Are there approved plans that match this request?
    • Return the three closest approved plans for this run.
read_agent_execution_memory – Returns the stored messages from a run, including the tool calls and reasoning behind the result.
  • When to use
    • Review the tools an agent called during a run.
    • Understand why an agent chose one approach over another.
    • Review how a workflow decided the input values for a step.
    • Trace the exchange that led to an unexpected result.
    • Page through the history of a long run.
  • Parameters
    • memory_id – The ID of the memory to read. Run and step records return this ID.
    • (Optional) start_sequence – The position in the message list where reading starts. Use the marker from the previous page to continue.
    • (Optional) limit – The number of messages to return. Defaults to 512, which is also the maximum.
  • Example prompts
    • Read the memory for execution exec-3390 and list the tools it called.
    • Show me the reasoning behind that workflow run.
    • What did the agent consider before it returned that answer?
    • Show me the next page of messages for that run.
search_agents – Finds specialized agents and workflow agents in your Opal instance by handle, name, or description.
  • When to use
    • Find an agent when you remember only part of its name.
    • List every agent available in your instance.
    • Check whether an agent exists before you build one.
    • Find the handle of an agent so you can run or edit it.
    • Discover agents that relate to a topic, such as email or campaigns.
  • Parameters
    • (Optional) search_query – The text to match against the handle, name, and description of each agent. An exact handle ranks highest. Leave this value empty to list every agent.
  • Example prompts
    • List all the agents in my instance.
    • Do I have an agent for reviewing documentation?
    • Find agents related to email.
    • What is the handle for my case study agent?

Skills

create_or_edit_skill – Saves a reusable instruction as a personal skill or an organization skill, and edits or renames an existing skill.
  • When to use
    • Save a standing instruction from a conversation so Opal applies it later.
    • Change the content, title, or trigger settings of an existing skill.
    • Add a shortcut so you start a skill by name in chat.
    • Attach reference files, such as templates or brand guidelines, to a skill.
    • Share a personal skill with everyone in your organization.
  • Parameters
    • skill_type – The audience for the skill. Use personal for your own use, or org to share the skill across your organization. A change of audience requires a skill.
    • (Optional) skill_id – The ID of the skill to edit. Omit this value to create a skill.
    • (Optional) title – The display name of the skill, such as *Translate to Dothraki*. Required when you create a skill.
    • (Optional) skill – The instructions Opal follows when the skill applies. Write complete instructions, including which tools to use. Required when you create a skill.
    • (Optional) mention_handle – The shortcut that starts the skill in chat, entered without the slash. Use lowercase letters, numbers, and hyphens.
    • (Optional) trigger_type – How Opal activates the skill. Use keyword to activate on matching phrases, or tool to activate when specific tools are present.
    • (Optional) trigger_conditions – A plain-language description of when Opal applies the skill.
    • (Optional) trigger_tools – The tools that activate the skill when they are present in a session.
    • (Optional) active – Makes the skill available in chat. Skills you create are active by default.
    • (Optional) files – The files attached to the skill, such as a template or a script. Upload each file to the conversation first. Pass an empty list to remove every attached file.
    • (Optional) instance_id – The instance that the skill applies to. Include this value only when you limit the skill to one instance.
    • (Optional) product_sku – The product that the skill applies to. Include this value only when you limit the skill to one product.
    • (Optional) update_summary – A short summary of what changed, such as *Add shortcut*.
  • Example prompts
    • Save this as a skill.
    • Create a skill that formats my release notes as a table.
    • Add a waterways shortcut to my AmaWaterways skill.
    • Change the title of my newsletter skill to "Newsletter copywriter".
    • Share this skill with my whole organization.
find_relevant_skills – Returns the skills that best match a described intent, ranked by how well each skill fits.
  • When to use
    • Check whether a saved skill fits an open-ended request.
    • Find a skill by describing the work instead of naming the skill.
    • Search when you do not remember the title or the shortcut.
    • Review which personal skills and organization skills apply to a topic.
  • Parameters
    • context – The intent or the message that Opal matches skills against.
    • (Optional) limit – The number of skills to return, best match first.
  • Example prompts
    • Find any skills that apply to drafting a competitive research report.
    • Do I have a skill that fits summarizing a meeting into action items?
    • Check for saved guidelines that relate to this brand copywriting request.
find_skill – Finds a skill by matching text against the title, content, and shortcut of each skill.
  • When to use
    • Find a skill by name or by shortcut so you edit it.
    • Check whether a similar skill exists before you create one.
    • Find the ID of a skill for a later edit.
    • Limit a search to personal skills or organization skills.
    • Confirm whether a skill is active.
  • Parameters
    • query – The text to match against the title, content, and shortcut of each skill. A single distinctive word returns better results than a full phrase.
    • (Optional) kind – Limits the search to personal skills or organization skills. Omit this value to search both.
  • Example prompts
    • Find my standup skill.
    • Do I have a skill for writing newsletters?
    • Look up the skill with the waterways shortcut.
    • Show me my organization skills about documentation.
get_skill – Returns the full definition of one skill.
  • When to use
    • Read the content of a skill before you change it.
    • Review the exact instructions a skill applies.
    • Review the attached files before you add or remove one.
    • Confirm the trigger type and trigger conditions of a skill.
    • Document the configuration of an existing skill.
  • Parameters
    • skill_guid – The ID of the skill to return. A skill search returns this ID.
  • Example prompts
    • Show me the full content of that skill.
    • What does my documentation style skill say?
    • Read the skill before you change it.
    • Which files are attached to this skill?

Canvas and artifacts

Your active canvas provides a dynamic and collaborative environment where you and Optimizely Opal can build, customize, and deliver interactive digital assets, applications, and complex content. This workspace lets you create and manage digital assets in real-time. 

canvas_to_file – Exports a canvas to a file, including PDF for markdown canvases and HTML canvases.
  • When to use
    • Download a canvas to share it outside Opal.
    • Convert a markdown canvas or an HTML canvas to PDF.
    • Save a copy of a report before you edit the canvas further.
    • Produce a file to attach to an email.
    • Give the exported file a specific name.
  • Parameters
    • canvas_id – The ID of the canvas to export.
    • (Optional) content_type – The format of the exported file. Opal uses the format of the canvas when you omit this value.
    • (Optional) filename – The name of the file, including the extension. Opal generates a name when you omit this value.
    • (Optional) workspace_id – The space that holds the canvas. Opal detects the space when you omit this value.
  • Example prompts
    • Export this canvas as a PDF.
    • Download my campaign brief as q3-brief.pdf.
    • Save this report to a file so I can email it.
    • Turn the canvas into a PDF I can send to my team.
create_canvas – Creates a canvas, an editable document that you and Opal work on together in real time.
  • When to use
    • Draft a document you plan to revise, such as a brief, a report, or a blog post.
    • Build a web page, an email mockup, or another HTML layout.
    • Create a React application or a Next.js page.
    • Produce a chart or a data visualization you want to adjust.
    • Keep long content out of the chat so you edit it in one place.
  • Parameters
    • name – The name of the canvas.
    • mime_type – The format of the canvas. Use text/markdown for documents and notes, text/html for web pages and mockups, application/x-reactjs for React components, and application/x-nextjs for Next.js pages.
    • (Optional) content – The starting content of the canvas.
    • (Optional) workspace_id – The space that holds the canvas. Opal uses the space of the current conversation when you omit this value.
    • (Optional) workspace_name – The name of the space when Opal creates one. Defaults to the canvas name.
    • (Optional) space_description – The description of the space when Opal creates one.
  • Example prompts
    • Create a campaign brief for our fall product launch.
    • Write a blog post about A/B testing and put it in a canvas.
    • Build an HTML email mockup for our newsletter.
    • Make a React dashboard that charts our conversion rates.
create_csv_canvas – Creates a canvas that holds comma-separated values (CSV) data, either from files you supply or from data Opal already has.
  • When to use
    • Turn exported data into a table you review and edit in Opal.
    • Merge several CSV files into one table.
    • Save search results or report data as a working spreadsheet.
    • Build a table you plan to add rows to over time.
  • Parameters
    • title – The name of the CSV canvas.
    • (Optional) content – The CSV data to display. Use this value or csv_urls, but not both.
    • (Optional) csv_urls – The addresses of the CSV files to download and merge. Use this value or content, but not both.
    • (Optional) workspace_id – The space that holds the canvas. Opal uses the space of the current conversation when you omit this value.
    • (Optional) workspace_name – The name of the space when Opal creates one. Defaults to the canvas title.
    • (Optional) space_description – The description of the space when Opal creates one.
  • Example prompts
    • Put these campaign results into a CSV canvas.
    • Merge these two CSV files into one table.
    • Export my overdue tasks to a spreadsheet I can edit.
    • Create a CSV canvas called Q3 Metrics from this data.
create_powerpoint_canvas – Creates a canvas from a PowerPoint file so you view and change the deck inside Opal.
  • When to use
    • Open an existing deck in Opal to review it.
    • Prepare a deck for edits, such as slide removal or text replacement.
    • Pull a deck into a conversation so Opal reads the slide content.
    • Keep a working copy of a presentation alongside your other content.
  • Parameters
    • powerpoint_url – The address of the .pptx file to open.
    • canvas_name – The name of the canvas.
    • (Optional) workspace_id – The space that holds the canvas. Opal uses the space of the current conversation when you omit this value.
    • (Optional) workspace_name – The name of the space when Opal creates one. Defaults to the canvas name.
    • (Optional) space_description – The description of the space when Opal creates one.
  • Example prompts
    • Open this PowerPoint in a canvas.
    • Create a canvas from our quarterly business review deck.
    • Pull this deck into Opal so I can edit the slides.
edit_canvas – Applies changes to a canvas, either as a targeted edit or as a full replacement of the content.
  • When to use
    • Revise a section of a document without rewriting the whole canvas.
    • Replace the entire content of a canvas.
    • Apply feedback to a draft you created earlier.
    • Fix an error in a report or a code canvas.
    • Add a section to an existing document.
  • Parameters
    • element_id – The ID of the canvas to change.
    • base_commit_id – The version of the canvas the change applies to. Opal tracks this value so an edit does not overwrite a change someone else saved.
    • patch_operations – The changes to apply. Opal builds these changes from your request as either a targeted edit or a full replacement.
    • (Optional) workspace_id – The space that holds the canvas. Opal detects the space when you omit this value.
  • Example prompts
    • Change the headline in my canvas to something shorter.
    • Add a risks section to the campaign brief.
    • Rewrite the introduction so it leads with the customer benefit.
    • Replace the whole canvas with this updated draft.
edit_csv_canvas – Changes the data in a CSV canvas by adding, updating, deleting, or reordering rows and columns.
  • When to use
    • Add rows to a table as work progresses.
    • Correct values in specific rows.
    • Remove rows you no longer need.
    • Reorder columns or drop the ones you do not use.
    • Replace the whole table with a refreshed data set.
  • Parameters
    • element_id – The ID of the CSV canvas to change.
    • base_commit_id – The version of the canvas the change applies to. Opal tracks this value so an edit does not overwrite a change someone else saved.
    • operation – The change to make. Use append_rows to add rows, update_rows to change existing rows, delete_rows to remove rows, reorder_columns to set the column order, and replace_content to swap the whole table.
    • (Optional) rows_json – The row data to add or apply. Required for append_rows and update_rows.
    • (Optional) row_indices – The positions of the rows to change or remove, counted from zero. Required for update_rows and delete_rows.
    • (Optional) column_names – The columns to keep and the order to display them in. Required for reorder_columns.
    • (Optional) new_csv_content – The replacement table. Required for replace_content.
    • (Optional) workspace_id – The space that holds the canvas. Opal detects the space when you omit this value.
  • Example prompts
    • Add these three campaigns to my CSV canvas.
    • Delete the first two rows of the table.
    • Update row 5 with the corrected budget figure.
    • Reorder the columns so Status comes first.
    • Replace the table with this updated export.
get_canvas – Returns the content and details of one canvas.
  • When to use
    • Review the content of a canvas without opening it.
    • Ask Opal to summarize or analyze a document you created earlier.
    • Reference one canvas while you draft another.
    • Confirm that a canvas exists and check its format.
  • Parameters
    • element_id – The ID of the canvas to return.
    • (Optional) workspace_id – The space that holds the canvas. Opal detects the space when you omit this value.
  • Example prompts
    • Show me what is in my campaign brief canvas.
    • Summarize the report canvas I created yesterday.
    • Read my messaging document and pull out the key claims.
    • What format is that canvas saved in?
get_canvas_by_workspace – Lists every canvas in a space, with the name, format, and creation date of each one.
  • When to use
    • Review the content in a space.
    • Find a canvas when you do not remember its name.
    • Check what exists before you create a canvas.
    • Choose which canvas to edit among several.
  • Parameters
    • (Optional) workspace_id – The space to list. Opal uses the space of the current conversation when you omit this value.
  • Example prompts
    • List all my canvases.
    • What is in this space?
    • Show me everything I created in this conversation.
    • Do I already have a canvas for the fall campaign?
get_canvas_workspace – Returns the details of the canvas space attached to the current conversation.
  • When to use
    • Check the name and description of the space you work in.
    • Confirm that the current conversation has a space.
    • Review when the space was created and last changed.
  • Parameters
    • (Optional) workspace_id – The space to return. Opal uses the space of the current conversation when you omit this value.
  • Example prompts
    • What space am I working in?
    • Show me the details of this canvas space.
    • When was this space created?
present_widget – Displays a chart, a table, or a set of metric cards directly in the chat.
  • When to use
    • View analysis results as a chart instead of a block of text.
    • Compare values across a data set in a table.
    • Highlight key metrics as cards.
    • Review a visualization in chat before you move it into a report.
  • Parameters
    • title – The title displayed above the widget.
    • element_id – The ID of the widget to display. Opal produces this value when it builds the widget.
    • space_id – The space that holds the widget. Opal produces this value when it builds the widget.
  • Example prompts
    • Chart these results for me.
    • Show my campaign performance as a table in chat.
    • Display the conversion numbers as metric cards.
update_powerpoint_canvas – Changes a PowerPoint canvas by deleting slides, replacing text, copying a slide from another deck, or adding an image.
  • When to use
    • Remove slides from a deck.
    • Replace a term throughout a presentation, such as an old product name.
    • Copy a slide from one deck into another.
    • Place an image into a slide.
    • Adjust a deck without leaving Opal.
  • Parameters
    • element_id – The ID of the PowerPoint canvas to change.
    • base_commit_id – The version of the canvas the change applies to. Opal tracks this value so an edit does not overwrite a change someone else saved.
    • command – The change to make. Use delete_slide to remove slides, search_and_replace to change text, copy_slide to bring in a slide from another deck, and add_image_to_placeholder to place an image.
    • (Optional) slide_numbers – The slides to remove, counted from one. Required for delete_slide.
    • (Optional) search_text – The text to find. Required for search_and_replace.
    • (Optional) replace_text – The replacement text. Required for search_and_replace.
    • (Optional) source_ppt_url – The address of the deck to copy a slide from. Required for copy_slide.
    • (Optional) source_slide_number – The slide to copy from the source deck, counted from one. Required for copy_slide.
    • (Optional) image_url – The address of the image to add. Required for add_image_to_placeholder.
    • (Optional) slide_number – The slide that receives the image, counted from one. Required for add_image_to_placeholder.
    • (Optional) placeholder_text_regex – The text that identifies which shape the image replaces. Matching ignores capitalization. Required for add_image_to_placeholder.
    • (Optional) workspace_id – The space that holds the canvas. Opal detects the space when you omit this value.
  • Example prompts
    • Delete slides 3, 5, and 7 from this deck.
    • Replace every instance of our old product name with the current one.
    • Copy slide 2 from this other deck into my presentation.
    • Add our logo to the placeholder on slide 1.

Search and context

search_input_guardrails – Lists the input guardrails in your Opal instance and the trained versions of each one. An input guardrail is a rubric that screens a prompt before an agent acts on it. Opal returns only the guardrails that have at least one trained version.
  • When to use
    • Review which input guardrails exist in your instance.
    • Find a guardrail to apply to an agent.
    • Check which versions of a guardrail finished training.
    • Confirm the name of a guardrail before you change an agent.
    • Narrow a long list of guardrails to the ones that match a name.
  • Parameters
    • (Optional) query – The text to match against guardrail names. Matching ignores capitalization. Omit this value to list every guardrail.
  • Example prompts
    • List the input guardrails in my instance.
    • Which guardrails can I apply to an agent?
    • Find the guardrail with safety in the name.
    • Show me the trained versions of my content guardrail.
search_optimizely_docs – Searches official Optimizely product documentation and returns an answer with links to the source articles.
  • When to use
    • Look up how to complete a task in an Optimizely product.
    • Confirm how a feature works before you configure it.
    • Find API or SDK reference details.
    • Troubleshoot an error or an unexpected result.
    • Get an explanation of a concept, such as Stats Engine or mutual exclusion.
  • Parameters
    • query – The question or topic to look up. Include the product name to get more precise results.
  • Example prompts
    • How do I create a feature flag in Feature Experimentation?
    • What is Stats Engine, and how does it calculate significance?
    • Show me how to configure a content type in CMS (SaaS).
    • How do I set up a holdout group?
    • What does a sample ratio mismatch mean in my experiment results?
search_past_context – Searches earlier parts of a long conversation for a decision, a fact, a file, or a topic. Opal keeps recent messages close at hand, so use this tool to reach detail from further back in the thread.
  • When to use
    • Recall a decision made earlier in a long conversation.
    • Find which file you uploaded earlier in the thread.
    • Retrieve a value, a name, or a link mentioned much earlier.
    • Pick up work after a conversation runs long.
    • Confirm what you agreed on before you move forward.
  • Parameters
    • query – A description of what to find, written in plain language.
  • Example prompts
    • What did we decide about the campaign timeline?
    • Which file did I upload earlier in this conversation?
    • Remind me what we said about the launch date.
    • What was the final headline we agreed on?

If you use Opti ID, administrators can turn off generative AI in the Opti ID Admin Center. See Turn generative AI off across Optimizely applications.