OpenClaw

OpenClaw's own ask_user works well inside a live chat session, where there's a conversation to ask into. It doesn't help a job OpenClaw itself started on a schedule, with no chat session open and no one currently looking at it — that's the gap the pendnt OpenClaw skill fills.

1. Install the skill

Add the pendnt skill to your OpenClaw install (from ClawHub, or by pointing at the skill directly):

openclaw skill add pendnt

2. Configure the API key

export PENDNT_API_KEY="aio_..."
# or, in OpenClaw's own config/secrets file:
# pendnt_api_key: "aio_..."

3. Use it from a cron-triggered skill run

Inside a cron job's skill logic, call the pendnt tools exactly like any other OpenClaw skill tool — they resolve to the same request_approval / notify / wait_for_event / schedule_wakeup primitives documented in the MCP reference, over the same remote MCP server (https://api.pendnt.dev/mcp) or plain REST calls if you'd rather not add an MCP dependency to a particular skill.

# pseudocode inside an OpenClaw cron skill
result = pendnt.request_approval(
    title="Send the weekly digest to the customer list?",
    details=f"{subscriber_count} recipients, draft attached",
    timeout_s=21600,   # 6 hours — the job can exit and resume later
    wait_s=25,
)

if result.status == "pending":
    # exit now; OpenClaw's cron will re-run this skill later, or
    # schedule_wakeup + wait_for_event can resume it without a full re-run
    pendnt.schedule_wakeup(in_s=1800, payload={"request_id": result.id})
    return

if result.status == "approved":
    send_digest()
else:
    pendnt.notify(message="Digest send was denied or expired", level="warning")

When to use this vs. ask_user

Use both if you like: ask_user for anything asked during a live OpenClaw conversation, the pendnt skill for anything a scheduled job needs to ask when nobody's in a session to hear it.