> ## Documentation Index > Fetch the complete documentation index at: https://docs.plurai.ai/llms.txt > Use this file to discover all available pages before exploring further. # Suite and traces stages > Read the Suite stage of the report, from the agent flow graph to every scenario and eval, and the Traces stage with the production failures the suite replays. The **Suite** stage (`#suite`) shows what the simulation tests your agent on: the routes through your agent's flows, every scenario that walks them, and every eval that scores them. The **Traces** stage (`#traces`), when you grounded the suite in production traces, shows the real failures the suite replays. ## Your test suite *The scenarios built to walk the routes that matter through the Agent's flows, and the Evals that score what it must get right in them.* ![The Suite stage](https://mintcdn.com/plurai/V3XeGRFO5iyJejs5/images/report/suite.png?fit=max\&auto=format\&n=V3XeGRFO5iyJejs5\&q=85\&s=3634b5b3ffcfc88466a5ff8d6da88a8b) The Suite stage: select a scenario and its route lights up on your agent flow graph. The header counts the **Scenarios** and the **Evals**. The **Suite** stage tab shows the same counts. ### While the suite is being built The section fills in step by step. It shows a **Draft** note until the suite is final. | You see | Meaning | | - | - | | **Calibration** scenarios | *These scenarios are calibrating the simulator. The real Suite replaces them when it's ready.* Don't read them as the suite. | | *Scenarios landed · Evals still being written* | Scenarios exist. Evals are still being generated. | | *Scenarios and Evals landed · the Suite is still being built* | Both exist. The quality loop is still checking them. | ### Knowledge coverage When your agent answers from documents, the header adds a line on how much of the [knowledge base](https://docs.plurai.ai/report/scan.md#your-knowledge-base) the suite tests, for example *5 planted questions touch 9 of 214 passages · 1 of 3 documents never tested*. A document that is never tested is one the suite asks nothing about. **See it on Scan** opens the knowledge base. ![Tested knowledge](https://mintcdn.com/plurai/V3XeGRFO5iyJejs5/images/report/tested-knowledge.png?fit=max\&auto=format\&n=V3XeGRFO5iyJejs5\&q=85\&s=467773b8deb015ba96963e1ed1ae1faa) Tested knowledge: which facts from your knowledge base the suite checks. ## Your agent flow graph *Analysed agent modes and flows. Synthetic scenarios are based on these flows to keep them authentic and cover the routes that matter.* Below the header, the graph draws your agent's flows and the parts it uses along the way. The scan maps the flows from your code; the routes land on them as the suite build writes each scenario. | Element | What it is | | - | - | | **Tools** | Functions your agent calls. The graph draws them as the agent *calls tools*. | | **Skills** | Skills your agent *invokes*. | | **Sub-agents** | Agents your agent *delegates* to. | | **RAG** / **Knowledge** | Retrieval sources that *ground* an answer. | | Route | The path one scenario walks through the flows, for example `intake → lookup → refund`. | | `12% of conversations walk this route` | How often the route is expected, when the scan could estimate it. | | `×2` | The scenario walks this step more than once. | ### Read the graph 1. **Select a scenario to light its route.** The rest of the graph dims, so you see which flows that scenario exercises. 2. Select a flow to filter the scenario list with **Show only scenarios that walk** that flow. The page shows how many scenarios walk it, for example `3 of 7 walk refund`. Use **Show all scenarios** to clear the filter. 3. Look for flows no scenario walks. Those are parts of your agent the suite does not test yet. Before the scenarios exist, the graph shows the flows only, with the note **Routes appear as the scenarios land**. If a flow is missing or wrong, say so in chat, for example *"the refund flow also calls `check_fraud`"*. The simulation redoes the scan and every stage after it. See [Shape your simulation](https://docs.plurai.ai/guides/iterate.md). ## Scenarios Each scenario is a card. Select one to open its drawer, or link to it with `#scenario-`. | Field | Meaning | | - | - | | **Goal** / **User goal** | What the simulated user is trying to get done, in their words. | | **Who** | The [persona](https://docs.plurai.ai/concepts.md) who plays the scenario. | | **Speaks** | How the persona talks: tone, language, style. | | **Opens with** / **Opening line** | The first thing the persona says. | | **Passes when** | What a good outcome looks like for this scenario. | | **Scenario target** | The specific behavior the scenario is built to test. | | **Premise** / **The plan** | The setup and the beats the conversation is expected to go through. | | **Twists** / **Dealt in** | Complications built into the scenario, and the variations dealt into it. | | **Files the user brings** | Documents the persona hands the agent, such as an invoice or a PDF. | | **Questions from your documents** | Knowledge questions planted in this scenario, each with **Expected** answer and **Answer in** (the passage that holds it). | | **Route** | The path the scenario walks through your agent's flows. | | **Evals** | Which evals score this scenario, for example *All 6 of the Suite's Evals score this scenario*. | | **Provenance** | Where the scenario came from. See the next table. | | Provenance | Meaning | | - | - | | **Planned from the PRD** | Generated from the simulation PRD the scan wrote for your agent. | | **Replays a production failure** / **From production** | Built from a failure in your production traces. It continues the real conversation. | | **A Regression** / **Regression** | *Replays the turn a Gap was confirmed on, so it's caught if it comes back.* See [Gaps and fixes](https://docs.plurai.ai/report/gaps-and-fixes.md#kept-as-regressions). | | **Written for this Suite** | Written for this suite, not drawn from the PRD, production traces or a regression. | ## Evals Each eval is a row in **Evals**. Select one to read it, or link to it with `#eval-`. | Field | Meaning | | - | - | | Name and scope | What it checks, scored **over the whole conversation** or **on every turn**. | | Severity | How bad a failure of this eval is: **Critical**, **High**, **Medium**, **Low**, **Advisory** or **Info**. | | **Fails when** | The condition that fails the conversation on this eval. **Also fails when** lists more conditions. | | **Judged by** | Who scores it: **a model judge**, **code**, **knowledge-question grading** or **a retrieval measure**. | | Scoring | **Pass or fail**, **Pass or fail, at ≥ 0.8 confidence** for a model judge, or **Score** with its threshold. | | **Comes from** | Why the eval exists, with a quote: the **PRD**, **You asked** (your words in chat), a **Tool**, a **Scenario goal** or an **Eval brief**. | To change an eval, say so in chat or run `/edit-evals`. To change a scenario, see [Shape your simulation](https://docs.plurai.ai/guides/iterate.md). ## Production traces *Failures seen in production conversations, and which of them the Suite replays.* ![The Traces stage](https://mintcdn.com/plurai/V3XeGRFO5iyJejs5/images/report/traces.png?fit=max\&auto=format\&n=V3XeGRFO5iyJejs5\&q=85\&s=29d9d4586b99c6b56de7e9db168229be) The Traces stage: the real failures found in your production traces. The **Traces** stage (`#traces`) holds one section. It fills in when you run [Ground in production traces](https://docs.plurai.ai/guides/ground-in-production-traces.md) (`/analyze-traces`). The header says what the suite was built against, for example *Built against 240 production conversations from LangSmith.* | Count | Meaning | | - | - | | **production conversations** | How many real conversations were analyzed. | | **seen in production** | How many distinct failures were found in them. | | **replayed in the Suite** | How many of those failures have a scenario in the suite. | Each failure under **Seen in production** shows: | Label | Meaning | | - | - | | **Seen in** | How many conversations it appeared in, and how many times. | | **caught by** | The evals that catch it. **caught by no Eval** means no eval in the suite checks for it yet. | | **origin** / **fault** | Where the failure starts and which part is at fault, when the analysis could tell. | | **Replayed in the Suite** | Opens the scenario that replays this failure. | | **Not in this Suite** | The failure has no scenario in this suite. | | **Held back** | A scenario was planned but kept out, with the reason, for example *it did not fit under the scenario cap*. | | **Didn't ship** | A scenario was attempted and failed, with the reason, for example *authored, and did not pass validation*. | If you didn't ground the suite, the **Traces** stage reads *Skipped: this Suite isn't grounded in production traces. To add them, run `/analyze-traces` in Claude Code.* If the traces held no failures, the section says *No production failures were found in these traces.* ## Next steps > [card] [Run results](https://docs.plurai.ai/report/run-results.md) > > How each scenario played and how the evals scored it. > [card] [Build the suite](https://docs.plurai.ai/guides/build-the-suite.md) > > How the suite is generated and checked before it runs. This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.