desk-monitoring
Skill0 starsChanged 2 months agoMITView on GitHub
How the desk keeps an eye on markets and the account between trades - scheduled routines, the daily desk brief, book checks, and watches built on polling or the WebSocket feeds - including what a watch is allowed to do and why silence from one is never an all-clear. Use when the user asks for briefings, alerts, "watch X", or scheduled checks.
2 other projects use a version of this, 1 of them edited.
Adopted by
Most recently taken up by
- galleonlabs/hypergrok-trading-deskedited, first seen Nov 12, 2026
- swcstudiospace/PumpGroktheir own version, first seen Nov 5, 2026
Other versions2 projects use a version of this.
Word for word · 0
None found.
Edited · 1
- galleonlabs/hypergrok-trading-desk★ 53021% same
Wrote their own version · 1
- swcstudiospace/PumpGrok★ 66same name
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository22 other files it gives its agents.
- AGENTS.mdAGENTS.md
- desk-autopilotSkill
- desk-execution-protocolSkill
- desk-incident-responseSkill
- desk-operating-modelSkill
- desk-post-trade-reviewSkill
- desk-risk-limitsSkill
- desk-signal-scanSkill
- desk-standing-approvalsSkill
- desk-strategy-labSkill
- desk-trade-lifecycleSkill
- strike-accountSkill
- strike-advancedSkill
- strike-api-referenceSkill
- strike-authSkill
- strikegrok-bootstrapSkill
- strike-market-dataSkill
- strike-ordersSkill
- strike-positionsSkill
- strike-research-toolsSkill
- strike-setupSkill
- strike-websocketSkill
The file
We split this skill into the rules it gives an agent
7 rules. About 0 are common, 6 are distinctive, and 1 is very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈6
Distinctive · estimate
This author's own choices. Worth a look.
1
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 83
Every watch logs to /workspace/trading-desk/watch/<name>/log with UTC timestamps, and keeps its condition written in plain language beside …
- Distinctive: line 85
Run a few, not many. Each is a process on a shared computer.
- Distinctive: line 87
When the desk goes flat for the day, stop the watches that only make sense alongside a position.
- Distinctive: line 91
Never let a watch place, change or cancel an order, or touch leverage — not even a protective one.
- Distinctive: line 92
Never alert with a number that has no source.
- Distinctive: line 93
Never let a stale or failed read pass as a condition that did not fire.
- Distinctive: line 94
Never poll faster than the desk needs.
≈ “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
Monitoring, briefs and watches
Everything here reads. A watch may fetch, compute, save and alert; it may not touch the write path. When a condition implies an action — "the stop went, get back in" — the alert opens a proposal and the lifecycle takes it from there.
Routines
Each Bot can hold scheduled or triggered routines. Ask the owning Bot for one and specify four things: when it runs and in which time zone, what it reads, what it should produce, and where its authority stops. Keep them few and specific. The platform retains only recent run records, so anything meant to last should also write into /workspace/trading-desk.
A reasonable starting set — the user picks, none runs by default:
| Routine | Owner | When | Produces |
| Desk brief | Desk Lead | daily, at a time the user picks | briefs/YYYY-MM-DD-desk.md plus a short post |
| Book check | Risk Manager | every 4–8 hours while anything is open | an alert only when something moved or a limit is close |
| Funding snapshot | Market Analyst | daily, before the user's usual hours | funding, open interest and volume across the allowed markets |
| Calendar refresh | Research Analyst | weekly | research/calendar.md |
| Weekly review | Trade Reviewer | weekly | the review, by DM |
The desk brief
The Desk Lead assembles it from the specialists, or produces it alone if it is the only Bot awake.
DESK BRIEF | 2026-09-11 07:00 UTC
markets (Market Analyst, /v2/ticker/24hr + /v2/premiumIndex 06:58 UTC)
BTC-USD 77,120 +0.8% 24h funding +0.0079%/h OI $682K
ETH-USD 2,431 +0.2% funding +0.0024%/h OI $132K
ADA-USD 0.2116 -1.1% funding -0.0004%/h OI $249K
book (Risk Manager, GET /v2/account + /v2/positions 06:59 UTC)
1 position: ADA-USD long 4800 @ 0.2100, uPnL +$7.68, margin ratio 4.6%
stop resting at 0.1995 (status 5, untriggered) | open risk 0.5% | day PnL 0.0%
research (Research Analyst)
NVDA earnings 2026-09-17 after the close [link]; nothing breaking on held markets
open items
SG-20260910-01 live; no pending ticketsFacts carry sources and times. Interpretation, if there is any, goes on one line that says it is interpretation. No trade suggestions.
Watches
A watch is a condition plus an alert. Two ways to run one from the desk computer:
Polling. A loop in a script under /workspace/trading-desk/watch/<name>/, hitting the public Price Service. Mind the weight limit — 2400 per minute — but for a desk, one read every five to fifteen seconds per market is ample.
Streaming. Subscribe to wss://api.strikefinance.org/ws/price for mark price, depth, trades and klines, or the authenticated user stream for fills, orders and positions. Better when the condition is genuinely sub-minute or the desk needs fills the moment they land. Keep the process supervised and logged. See strike-websocket.
Conditions worth running:
| Watch | Read | Tell |
| Price crosses a level | /v2/markPrice, or klines | user, Desk Lead |
| Funding flips sign or passes a threshold | /v2/premiumIndex | user, Desk Lead |
| An order fills or cancels | user stream, GET /v2/history/fill | Execution Trader, Trade Reviewer |
| Margin ratio above X, or liquidation closer than Y | GET /v2/account, GET /v2/positions | user, Risk Manager, Desk Lead |
| A position with no resting stop | GET /v2/positions + GET /v2/openOrders | Risk Manager, Desk Lead — incident |
| Daily loss stop at 80%, or hit | GET /v2/account against start-of-day equity | user, Risk Manager |
| A TWAP still running | GET /v2/algo/twap/{id} | Execution Trader |
| The exchange unreachable beyond N minutes | any read failing | Desk Lead |
An alert states what fired, the value against the threshold, the source and UTC time, and the proposal id where there is one. It never carries an instruction to trade.
Silence is not an all-clear
A watch has three possible outcomes, never two: the condition fired, the condition did not fire, or the watch could not tell. A failed request, an empty response, a socket that reconnected across a gap, a candle that never closed, or a value older than the watch's own interval all belong in the third category, and all must alert as such.
A watch that quietly reports "not crossed" over a dead feed is the most dangerous process on the desk, because it is indistinguishable from a calm market right up until it isn't.
So every watch gets a staleness bound — a small multiple of its interval — and checks each result's age against it. When it trips, the alert says which read failed, how old the last good value was, and how long the gap has run. A gap that outlives the bound escalates rather than waiting politely for the feed to return.
Hygiene
- Every watch logs to
/workspace/trading-desk/watch/<name>/logwith UTC timestamps, and keeps its condition written in plain language beside it. - A watch that dies quietly is worse than no watch. Whatever owns it checks its heartbeat: uptime, last message, reconnect count, and whether it currently believes itself live.
- Run a few, not many. Each is a process on a shared computer.
- Record running watches in
desk.md. One nobody remembers starting is one nobody notices stopping. - When the desk goes flat for the day, stop the watches that only make sense alongside a position.
Never
- Never let a watch place, change or cancel an order, or touch leverage — not even a protective one. It alerts; the desk acts through a ticket.
- Never alert with a number that has no source.
- Never let a stale or failed read pass as a condition that did not fire.
- Never poll faster than the desk needs. A rate-limited desk is a blind desk.
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.

