Dobeu-tech-eco

dt-cooldown

"Pause an agent that hit a usage limit until its reset date, resume it, or list what is paused. Use for set a cooldown, agent is out of usage, clear a cooldown, or which agents are paused."

Dobeu-tech-eco 0 Updated 1w ago
GitHub

Install

npx skillscat add dobeu-tech-eco/dts2-marketplace/dt-cooldown

Install via the SkillsCat registry.

SKILL.md

Cursor procedure

Find the plugin root: the nearest ancestor of this file that contains .cursor-plugin/plugin.json. The examples below say $PLUGIN for that directory.

Run python3 "$PLUGIN/scripts/dt_gate.py" for every gate. Exit 2 from scaffold, preflight, or approve stops the command. Exit 2 from bridge that prints BLOCKED unsupported-runtime-script is a precise blocker for that one step: record it and do not imitate the script. The interview, approve, and scaffold still run.

--yes, --ship, and --scaffold-only are rejected. A JSON key with the same name as a kind is not authorization. approve records the sha256 of the file the user accepted, and later gates require that digest.

Phase files for this command are in $PLUGIN/skills/dt-cooldown/.

Merges, fleet broadcast, hook install, the headless SDK, and paid provider calls are CURSOR_UNSUPPORTED even after a permission phrase is recorded.

Cursor edition: run scripts/dt_gate.py from this plugin for every gate. Do not read tokens or credential files. Do not merge, broadcast, install hooks, or spend a provider API. Carry out file work with Read, Edit, Grep, Glob, and git.

Cooldown — pause an agent until its usage limit resets

Usage limits belong to the account, so one status file (~/.dts2/agent-status.json, or the path in DTS2_AGENT_STATUS) is shared by every session on the machine. The roster probe, the lead failover and the SessionStart hook all read it. A paused agent is never called, probed or picked as lead until its reset time passes; it then comes back on its own.

All commands: python3 "$PLUGIN/scripts/dt_gate.py" bridge cooldowns.py <command>.

Steps

  1. set set <agent> <date or duration> [reason]
    • A duration such as 2d 3h runs set <agent> --in "2d 3h" --reason "<reason>".
    • A date or time such as 2026-11-21, Nov 21 9:00 AM or an ISO timestamp runs set <agent> --until "<date>" --reason "<reason>".
    • A date that cannot be read exits 2: ask for another form once.
  2. clear clear <agent> removes the pause now (the limit was lifted early, or the date was wrong).
  3. list list shows the agents paused now with their reset time and reason; list --all also shows entries that have ended.
  4. Confirm in one line: which agent is paused, until when (local time and UTC), and that it is skipped everywhere until then.

First agent to hit a limit

When a dispatch or a call fails with a usage, quota or rate-limit message:

  1. Pipe the output to detect --agent <agent> --exit-code <n> --record.
  2. If the result says "limited": false, this is not a usage limit; handle it as an ordinary failure.
  3. If it says needs_date: true, ask the user ONCE for the reset date or duration, then run set <agent> --until "<date>" --source user. A provisional pause of about five hours stays in place until they answer.
  4. Tell the user which agent is paused and until when, then continue the work with the remaining agents.
  5. Never re-probe or re-call a cooling agent. check <agent> exits 3 while it is paused; do not try to "see if it is back".

Notes

  • A limit that resets in under 15 minutes is not recorded; wait and retry instead.
  • A later, shorter detection never shortens a pause that is already active; only set (a person's word) does.
  • The SessionStart hook prints which agents are paused, and which came back, at the start of each session.