desk-tick
Skill36 starsChanged 3 days agoMITView on GitHub
One tick of a desk's trading loop on pathiel, for the desk id given as the argument (oil, metals, ai, gas). Slot check here, then the desk-tick agent (Opus 5.5, medium effort) runs sweep, levels, brief, one decision, one log line. Run on a schedule with /loop 15m /pathiel:desk-tick oil. Live money on the connected Hyperliquid account.
In this repository6 other files it gives its agents.
- CLAUDE.mdCLAUDE.md
- desk-loopsSkill
- desk-tickSkill
- event-calendarSkill
- review-tradesSkill
- pathiel-agentSkill
The file
We split this skill into the rules it gives an agent
2 rules. About 0 are common, 0 are distinctive, and 2 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈0
Distinctive · estimate
This author's own choices. Worth a look.
2
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 17
If slot is inactive, append DESK-TICK $ARGUMENTS <utc> action=SKIP desk not active to /tmp/oil-loop/progress.log, print it, and STOP.
- Very specific: line 47
If the agent fails to launch or returns no log line, append DESK-TICK $ARGUMENTS <utc> action=NONE agent did not run to the log and say so.
≈ “Common” is an estimate: at least 1 in 5 of the 372,429 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
<!-- GENERATED from .claude/skills/desk-tick/SKILL.md by scripts/build_plugin_skill.py. Edit the source, then rerun it. -->desk-tick $ARGUMENTS
One tick for desk $ARGUMENTS (default oil if empty). Two parts: the slot
check runs here because it is one script call, and the tick itself runs in
the desk-tick agent so the session's own model is not spent on a
15-minute loop (operator, 2026-10-25: "fable is valuable and the tick is
eating my usage"; an hourly clock was tried that evening and reverted the
same night).
- Slot first.
"${CLAUDE_PLUGIN_ROOT}/scripts/pathiel-py" scripts/desk_slot.py --desk $ARGUMENTS. Ifslotisinactive, appendDESK-TICK $ARGUMENTS <utc> action=SKIP desk not activeto/tmp/oil-loop/progress.log, print it, and STOP. Ifslotisheldandmineis false, another desk's coin holds the account's one slot: appendDESK-TICK $ARGUMENTS <utc> action=SKIP slot held by <holder>, print it, and STOP. No agent, no MCP calls. With desk equity shares saved (configdesk_equity_share; metals = gold since 2026-11-05) every desk owns its own slot, so the second case only happens without shares (the 2026-10-23 rule: the desk holding the one slot is the only one that ticks). - Otherwise launch the agent. Agent tool,
subagent_typepathiel:desk-tick, promptDESK: $ARGUMENTS. Do not run the tick steps yourself and do not launch more than one agent per tick. The agent file (the plugin'sagents/desk-tick.md) carries the whole procedure and its hard rules. Dead-tick rule (operator, 2026-11-09, after three stalls): a previous tick for this desk counts as running only if it started under 20 minutes ago. One older than that died with the network or a sleeping Mac and never reported back: do NOT skip for it. AppendDESK-TICK <desk> <utc> action=NONE previous tick presumed dead after <N> minto/tmp/oil-loop/progress.logand launch the agent now. This overrides the desk-loops overlap rule. - Relay, quietly. The agent already wrote its line to
/tmp/oil-loop/progress.log. Say NOTHING in the chat for a HOLD, a SKIP, or a NONE whose reason repeats that desk's previous log line (operator, 2026-11-02: "can u hide these messages, it crowds up the dialogue box"). Speak only when something changed: PLACED, CANCELLED, CLOSED, a stop or target fill, a NONE with a new reason, an uncovered position, or a tool error. Then one short message: what happened, with the numbers, and the log line.
If the agent fails to launch or returns no log line, append
DESK-TICK $ARGUMENTS <utc> action=NONE agent did not run to the log and
say so.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

