> ## Documentation Index > Fetch the complete documentation index at: https://docs.clarityq.ai/llms.txt > Use this file to discover all available pages before exploring further. # Context Coverage > See how complete your Context Layer was for each answer, and close the gaps it found. Every answer in Ask Anything carries a **context coverage** line telling you how solid the Context Layer was for that particular question. It reports on the *knowledge the answer used* — not on whether the answer is right. ## Reading the Badge The line sits under the answer, next to the rating buttons. | What you see | What it means | | - | - | | **Fully covered** (green) | Everything the answer used is defined and approved | | **Review N gaps** (orange) | The answer leaned on something not yet approved, or you corrected the agent in this chat | | **Review N gaps** (red) | Something the answer needed is missing from your Context Layer | Orange and red are clickable. Green isn't, because there's nothing to look at. The badge is **historical**: it records the state at the moment the answer was produced and never changes afterwards, even once you fix the gap. Later answers will show the improvement. ## The Context Coverage Report Clicking **Review N gaps** opens the **Context coverage report** beside the chat. You can also reach it from the chat's actions menu, which shows **Context coverage (N)** for the whole conversation — that number is deduplicated, so it's usually smaller than the badges added together. The report splits what it found into two groups: * **Items that require your approval** — the answer used something that's defined but not yet approved. * **Items to add/edit in context** — something is missing, needs updating, or a table changed underneath an entity. Each row names the item and explains the gap. Items that no longer need attention stay visible, marked **Resolved** or **Dismissed**, so an old answer's badge still makes sense when you look back at it. ## Acting on a Gap What you can do depends on your role. If you can edit the Context Layer: * **Review** — opens the item so you can **Edit**, **Approve**, or **Exclude** it without leaving the chat. * **Add in Builder** / **Edit in Builder** — opens the Builder with the change already described, ready for you to refine and build. * **Fix manually** — for [features](/context-layer/semantic-catalog/features) and their dimensions, which aren't built in the Builder. This takes you to the right page instead. * **Send to Inbox** — files a correction you made in the chat so someone can act on it later. If you can't edit the Context Layer, you get **Report** instead. It sends everything the conversation turned up to the [Inbox](/context-layer/building/inbox), along with a link to the chat so whoever picks it up can see the context. This is how you flag "the agent got this wrong" without needing to fix it yourself. Corrections you make in chat are never filed automatically. Telling the agent it's wrong helps that conversation, but it only reaches the Inbox when someone presses **Report** or **Send to Inbox**. ## In Shared Conversations A [shared conversation](/ask-anything/sharing-and-collaboration) shows the badges and the report, read-only. Nobody can approve, fix, or report from a shared view. The findings come from the snapshot taken when the chat was shared, so anything added afterwards isn't included. ## Where It Doesn't Appear Context coverage is for regular Ask Anything chats. It isn't shown in Builder chats or in an [embedded](/integrations/embed/embedding-clarityq) ClarityQ. This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.