design-the-rejection-path-too

Make an approval gate real — the draft visible where the record lives, a cap on what one approval covers, and a rejection path somebody has actually taken.

Procedure

  1. start start. Goes next to s1
  2. action List every action it can take, RO (read-only). Goes on_fail to s9; next to s2
  3. decision Which of them needs a person?. Goes if some_gated to s3; if default to s9
  4. check The draft looks like a draft downstream. Goes on_fail to s4; next to s5
  5. action Mark it unreviewed where it lands, RW (reversible-write). Goes on_fail to s9; next to s5
  6. decision How much does one approval cover?. Goes if uncapped to s6; if default to s7
  7. action Cap it, and split what exceeds the cap, RW (reversible-write). Goes on_fail to s9; next to s7
  8. action Answer the five rejection questions, RO (read-only). Goes on_fail to s9; next to s8
  9. action Reject something on purpose, end to end, RW (reversible-write), requires approval. Goes on_fail to s9; next to __end_success
  10. escalate The gate is not a gate yet. Terminal
  11. end success. Terminal

Open the graph on its own page — larger, printable, and an address you can send.

Walk it without running it — you answer what your environment would answer, and the path, the capabilities, the approvals and the distance back to an undo come out of the document. Nothing executes; there is no Run button here and there will not be one.

Trust

T2 · attested 2026-10-11 · runbooks.directory

Nobody has attested this record. That is the ordinary case: T2 is what the catalog can check on its own, and T3 needs somebody to run the procedure and sign for it.

  • Link liveness has not been checked: it needs a network and the catalog builds without one. Run pnpm gates --online where there is one.

Steps

  1. s1 action — List every action it can take (read-only)
  2. s2 decision — Which of them needs a person?
  3. s3 check — The draft looks like a draft downstream
  4. s4 action — Mark it unreviewed where it lands (reversible-write)
  5. s5 decision — How much does one approval cover?
  6. s6 action — Cap it, and split what exceeds the cap (reversible-write)
  7. s7 action — Answer the five rejection questions (read-only)
  8. s8 action — Reject something on purpose, end to end (reversible-write) — requires approval
  9. s9 escalate — The gate is not a gate yet

Capabilities

This is a privilege request. Check it against your own allowlist before running.

    Provenance

    Source
    https://github.com/microsoft/ai-agent-runbooks/blob/083f33eec889/02-patterns/Human-in-the-Loop-Review-and-Approval/Human-in-the-Loop-Review-and-Approval-Runbook.md
    Fetched
    2026-09-04
    Source last checked
    not since it was captured on 2026-09-04 — recorded as current
    Licence
    MIT
    How it got here
    Rewritten here from the source above, under its licence. The words and the steps are this catalog's; the procedure it describes is not this catalog's invention, and the source is where to check it.
    Highest risk in this procedure
    reversible-write

    Take it with you

    Run it under supervision

    Runs here, on your machine, under supervision: every step is proposed before it happens, an approval gate stops for you, and a failed expectation goes where the procedure says rather than on to the next step. Nothing is sent anywhere. To supervise an agent that is doing the work instead, runbooks hook sits in its tool-call path.

    npx @runbooks/cli run adapted/design-the-rejection-path-too@1.0.0 --as "$(whoami)" \
  --input destination=<the system the approved thing is written into>