# Pending Writes

## What this is

A Pending Write is a Protege-authored change that is waiting for human review before it is applied in Hookshot™. Pending writes are created when the write policy for a queue or Protege requires approval.

## When to use it

Use this page when you need to:

* Review changes a Protege wants to make
* Approve, reject, or edit a pending change
* Understand what happens when a pending write goes stale

## What you need first

* A workspace with Proteges that write to the Work Hub under an approval-required policy
* Permission to review pending writes (workspace owner, admin, or team lead)

## Steps

### 1. Find pending writes

Open **Work Hub** → **Decisions** to see pending writes awaiting review.

Each pending write shows:

* The Protege that authored it
* The target item and queue
* The operation type (create, update, comment, etc.)
* Whether the write will project to an external provider (Asana or Jira)
* The policy reasons why approval was required

### 2. Review the change

Open a pending write to see:

* **Before snapshot**: The current state of the item before the change
* **After snapshot**: What the item will look like if the change is applied
* **Edited payload**: If you modify the change before approving, the edited version appears here

{% hint style="warning" %}
Always review both the before and after snapshots before approving. The Protege may have made assumptions about the current state that are no longer accurate.
{% endhint %}

### 3. Approve, reject, or edit

| Action                | What happens                                                                                                                         |
| --------------------- | ------------------------------------------------------------------------------------------------------------------------------------ |
| **Approve**           | The change is applied to the item. If the queue has provider projection enabled, the change is also pushed to the external provider. |
| **Reject**            | The change is discarded. You can optionally provide a reason.                                                                        |
| **Edit then approve** | You modify the payload before approval. The edited version is applied instead of the original.                                       |

### 4. Handle stale writes

A pending write becomes **stale** if the underlying item has changed since the write was created. Stale writes are flagged because the before snapshot no longer matches the current state of the item.

When you encounter a stale write:

* Review whether the change is still appropriate given the new item state
* If it is still valid, approve it (the system will apply it against the current state)
* If it is no longer valid, reject it

### 5. Reopen a rejected write

If you rejected a write by mistake, you can **reopen** it for re-review. The write returns to `pending` status and can be approved or rejected again.

### 6. Understand write policy and pending writes

Pending writes are created when the write policy evaluates to `approval_required`. The policy is determined by four sources:

1. **Work Hub default** — the global setting
2. **Queue policy** — per-queue override
3. **Protege policy** — per-Protege setting
4. **Sensitivity** — the operation type

If any source says `approval_required`, a pending write is created instead of applying the change directly.

Operations that always require approval:

* Approval-sensitive work
* Decision requests
* Provider projections (writes that will push to Asana or Jira)

### 7. Review provider impact

When a pending write targets a queue that has provider projection enabled, approving the write will also push the change to the external provider. The pending write view shows whether the write has provider impact so you can consider the external consequences before approving.

## How to verify

* Pending writes appear for Protege changes that require approval
* Approved writes are applied to items and reflected in the item detail view
* Rejected writes are discarded and not applied
* Stale writes are flagged with a clear indicator
* Provider-projected writes update the external provider after approval

## Common failures

* Approving a write without checking the after snapshot
* Ignoring stale indicators when the item has changed
* Approving a provider-projected write without considering the external impact
* Rejecting a valid write that could have been edited instead
* Assuming direct mode is safe for all queues without reviewing the operations

## Next step

* [Work Hub](/event-feed/work-hub.md)
* [Audit](/event-feed/audit.md)


---

# Agent Instructions: Querying This Documentation

If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.

Perform an HTTP GET request on the current page URL with the `ask` query parameter:

```
GET https://gitbook.tryprotege.com/event-feed/pending-writes.md?ask=<question>
```

The question should be specific, self-contained, and written in natural language.
The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.