For agents and technical readers

Inspect the governance Inbox and answer the right item

Read the caller-scoped attention worklist, inspect each original request and report only the outcome actually observed.

Contents

Read without widening authority

Call orient and discover current tool availability. Call get_governance_inbox:

{"window":"7d"}

Supported windows are 7d, 30d, 90d; the field is optional. The read composes existing authorised surfaces and adds no new permission. Inspect completeness/error indicators. A denied, degraded or incomplete read is not an empty Inbox.

Read the original object

For a held file proposal use the review/proposal tools advertised by the connection and the correction procedure. For a loop wait, read get_loop and get_case with returned identifiers. For an overdue decision/policy, use get_record. A loop replacement is not the same object as a file proposal.

Answer the exact request

Where act and the relevant action are exposed and authorised, a question-back card's qb_id is answered with this JSON-string envelope:

{"action":"queue.answer","item_id":"<returned qb_id>","answer":"{\"chosen\":\"<one exact option from the card>\"}"}

Do not put a bare option such as approve in answer; it is not the required envelope. Batch-bound cards have a different row-decision contract; follow their current schema rather than applying this simple question envelope to them.

A human approval request can require human_confirmation containing the person's exact reply and the request's current content hash. Show the request and obtain that reply; never compose it yourself. No guide grants authority to approve.

Check what continued

Read the original request after answering, then inspect the continuing case or saved source. An accepted answer is not a completed workflow. After a lost response inspect these objects before retrying; keep the original item reference.

If the connection omits act or excludes the required action, hand the item to the authorised person in the Portal or a compatible client and report the handoff, not a completed decision.

Practical guides for people and agents

Start with one useful job

Join the waitlist

Plain text for AI agents. Same content as the human page.

## For agents

- What Mensara is: it gives a team its own Mind, one shared memory for the team and their AI agents, built by Xanadu Labs LLC.
- Status: signup is not open. A person joins the waitlist at https://app.mensara.ai/#waitlist. Agents cannot create an account or sign up.
- How you can help your person: explain what Mensara does, check whether it fits a job their team repeats, and share the waitlist link.
- Agent references: https://app.mensara.ai/llms.txt lists every page as Markdown and the tool input references for connected agents.

---

# Inspect the governance Inbox and answer the right item

Operating procedures · For agents and technical readers · [How it works](https://app.mensara.ai/how-it-works)

Read the caller-scoped attention worklist, inspect each original request and report only the outcome actually observed.

[Page as Markdown](https://app.mensara.ai/how-it-works/review-attention-procedure.md)

[Human guide](https://app.mensara.ai/how-it-works/review-attention)

## Use the released interface with current authority

Availability: Core0.20.11 / Portal0.3.47 source checked; discover the actual installed client and enabled features.

Who can do this: The connected principal with the permissions named in this procedure

What changes: The procedure explains its actual reads, writes or external effects before action.

Limits: An input accepted by a source schema is not a successful installed-runtime task.; Keep human decisions, client exclusions and pending results explicit.

Verification: Source checked and independently reviewed; scoped runtime receipts are recorded separately.

## Read without widening authority

Call `orient` and discover current tool availability. Call `get_governance_inbox`:

```
{"window":"7d"}
```

Supported windows are `7d`, `30d`, `90d`; the field is optional. The read composes existing authorised surfaces and adds no new permission. Inspect completeness/error indicators. A denied, degraded or incomplete read is not an empty Inbox.

## Read the original object

For a held file proposal use the review/proposal tools advertised by the connection and [the correction procedure](https://app.mensara.ai/how-it-works/correct-memory-procedure). For a loop wait, read `get_loop` and `get_case` with returned identifiers. For an overdue decision/policy, use `get_record`. A loop replacement is not the same object as a file proposal.

## Answer the exact request

Where `act` and the relevant action are exposed and authorised, a question-back card's `qb_id` is answered with this JSON-string envelope:

```
{"action":"queue.answer","item_id":"<returned qb_id>","answer":"{\"chosen\":\"<one exact option from the card>\"}"}
```

Do not put a bare option such as `approve` in `answer`; it is not the required envelope. Batch-bound cards have a different row-decision contract; follow their current schema rather than applying this simple question envelope to them.

A human approval request can require `human_confirmation` containing the person's exact reply and the request's current content hash. Show the request and obtain that reply; never compose it yourself. No guide grants authority to approve.

## Check what continued

If the connection omits `act` or excludes the required action, hand the item to the authorised person in the Portal or a compatible client and report the handoff, not a completed decision.

## Related

- [Review what needs your attention](https://app.mensara.ai/how-it-works/review-attention)
- [Read a loop case and reconcile a wait](https://app.mensara.ai/how-it-works/inspect-loop-case-procedure)
- [Read, correct and verify memory](https://app.mensara.ai/how-it-works/correct-memory-procedure)

[Join the waitlist](https://app.mensara.ai/#waitlist)