Human-in-the-loop review workflow. The agent works on a multi-step
task autonomously, but after each step sits idle with a phase
marker instead of halting, waiting for the manager to send one of
approve, {:revise, hint}, or finish.
When to reach for this
You want an agent to produce draft work in steps, but a human (or a reviewing agent) needs to gate each step before the next one starts. Writing a document one paragraph at a time with review between paragraphs. Producing a release plan and approving each stage. Generating PR descriptions that need human sign-off before posting.
The critical design decision is not to halt. If you halt with
{:halt, state} to pause for review, you freeze the mailbox --
which means a subsequent {:prompt, ..., state} return from
handle_event/2 will silently enqueue but never dispatch. Idle
with a phase marker is the correct primitive for "wait for next
input from outside."
What it exercises in gen_agent
- Idle-with-phase-marker as a pause primitive.
handle_response/3returns{:noreply, state}withphase: :awaiting_reviewinstead of{:halt, state}. The agent is idle and its mailbox is live, so a subsequentnotify/2can dispatch the next turn. handle_event/2returning{:prompt, text, state}as a resume primitive. Each review decision produces the next prompt and transitions back to:drafting.- Multiple decision outcomes from the manager: approve (continue to next step), revise (redo with feedback), finish (halt).
- Terminal halt only happens on final approval or explicit finish -- not on intermediate pauses.
The pattern
One callback module. The manager's review decisions come in as
{:review, :approve | {:revise, feedback} | :finish} notifies.
defmodule Checkpointer.Agent do
use GenAgent
defmodule State do
defstruct [
:task,
:draft,
:current_step,
:total_steps,
phase: :drafting,
history: [],
feedback: nil
]
end
@impl true
def init_agent(opts) do
state = %State{
task: Keyword.fetch!(opts, :task),
current_step: 1,
total_steps: Keyword.get(opts, :total_steps, 3)
}
system = """
You are a writing assistant working on a multi-step task.
Each turn, produce one refinement of the current draft. Keep
each output concise. No preamble, no explanations -- just
the draft.
"""
{:ok, [system: system, max_tokens: Keyword.get(opts, :max_tokens, 300)], state}
end
@impl true
def handle_response(_ref, response, %State{} = state) do
draft = String.trim(response.text)
new_history =
state.history ++ [%{step: state.current_step, draft: draft, feedback: state.feedback}]
new_state = %{
state
| draft: draft,
history: new_history,
feedback: nil,
phase: :awaiting_review
}
# NOT {:halt, state} -- halt would freeze the mailbox and
# block handle_event's {:prompt, ...} return. Idle with a
# phase marker is the correct pause primitive.
{:noreply, new_state}
end
# --- Review decisions from the manager ---
@impl true
def handle_event({:review, :approve}, %State{phase: :awaiting_review} = state) do
cond do
state.current_step >= state.total_steps ->
{:halt, %{state | phase: :done}}
true ->
next_step = state.current_step + 1
prompt = next_step_prompt(state.task, state.draft, next_step, state.total_steps)
{:prompt, prompt, %{state | current_step: next_step, phase: :drafting}}
end
end
def handle_event({:review, {:revise, feedback}}, %State{phase: :awaiting_review} = state)
when is_binary(feedback) do
prompt = """
Revise the current draft based on this feedback: #{feedback}
Current draft:
#{state.draft}
"""
{:prompt, prompt, %{state | feedback: feedback, phase: :drafting}}
end
def handle_event({:review, :finish}, %State{phase: :awaiting_review} = state) do
{:halt, %{state | phase: :done}}
end
def handle_event(_other, state), do: {:noreply, state}
# --- Prompts ---
defp next_step_prompt(task, previous_draft, next_step, total) do
"""
You are on step #{next_step} of #{total} for the task: #{task}
Previous draft:
#{previous_draft}
Produce the next refinement. Each step should improve on
the previous -- tighter, clearer, more specific.
"""
end
endUsing it
{:ok, _pid} = GenAgent.start_agent(Checkpointer.Agent,
name: "pitch",
backend: GenAgent.Backends.Anthropic,
task: "a single-sentence elevator pitch for a time-tracking app",
total_steps: 3
)
# Kick off the first step.
{:ok, _ref} = GenAgent.tell("pitch",
"Write an initial draft for: a single-sentence elevator pitch for a time-tracking app")
# Wait for phase: :awaiting_review and read the draft.
# (Use a small poll loop or a helper in your manager module.)
%{agent_state: %{draft: draft, current_step: step}} = GenAgent.status("pitch")
IO.puts("step #{step}: #{draft}")
# Decide what to do next.
GenAgent.notify("pitch", {:review, :approve})
# ... or
GenAgent.notify("pitch", {:review, {:revise, "make it more specific about the target user"}})
# ... or
GenAgent.notify("pitch", {:review, :finish})
# After approve, the next step dispatches automatically. Loop:
# wait_for_review -> inspect -> decide -> repeat.
GenAgent.stop("pitch")Variations
- Multi-reviewer sign-off. Instead of a single
:approvecommand, require N distinct reviewers to each send a{:review, :approve, reviewer_id}before advancing. Track approvals in state, advance when the set is full. - Time-boxed review. If no review decision arrives within a deadline, auto-approve or auto-finish. Use a state timeout or an external watchdog that fires a notify.
- Branching plans. Instead of a linear step counter, store a tree of planned steps and let the reviewer choose which branch to explore next via a more complex notify shape.
- Diff-based review. For patterns where each step produces a
file change rather than a prose draft, replace
draftwith a proposed diff and let the reviewer approve/reject it. Commits happen viapost_turn/3once approved. See Workspace for the workspace plumbing half of that.