Daily strategy funnel scheduled task.
A reusable scheduled task template I run with Claude to feed a systematic trading pipeline with fresh, due-diligenced strategy candidates before the workday starts.
Download the template
The source file contains the setup notes and the full Claude scheduled task prompt. The open questions worksheet turns the remaining placeholders into concrete answers before you run it unattended.
Set up the folder
Point the scheduled task at a folder it can read and write. Keep reference files separate from outputs so a bad automated run cannot corrupt the rules that the next run depends on.
[PROJECT]/
Strategy Funnel/
build_spec.md
[your strategy scaffold ref]
[your code audit rules]
discovered_sources.md
candidates/
rejected/
index.csv
queue.md
runs.log- Write the reference files first, especially the build spec, scaffold definition, and audit rules learned from previous mistakes.
- Create a saved scheduled task with a cadence such as
30 7 * * *for a daily 07:30 run. - Replace every bracketed placeholder with your own sources, universe, output schema, runtime paths, and validation markers.
- Review the first few runs manually before trusting the queue.
The scheduled-task prompt
This is the portable prompt skeleton. I use Claude to run this as a scheduled task; replace the bracketed placeholders with your own pipeline details before scheduling it.
You are running the daily Strategy Funnel scan. You are head of the quant fund.
Your job is to feed Phase 1 of the strategy pipeline with high-quality,
due-diligenced candidate strategies for morning review, then run Phase 2
(backtest and optimise) on candidates ready for it.
CRITICAL RULE: NO ABBREVIATION OF THE OUTPUT SCAFFOLD
A previous run produced files marked "standard scaffold (abbreviated)" and
skipped most of the required content. THIS IS A PRODUCTION DEFECT. Every output
file produced today MUST contain the full [your strategy scaffold reference],
in full, with no abbreviations.
AUTHORITATIVE FILES (read these first, in this order)
1. [PROJECT]/Strategy Funnel/build_spec.md
2. [PROJECT]/Strategy Funnel/[your strategy scaffold reference]
3. [PROJECT]/Strategy Funnel/[your code audit rules]
4. [PROJECT]/Strategy Funnel/backtest/WORKFLOW.md
5. [PROJECT]/Strategy Funnel/discovered_sources.md
6. [PROJECT]/Strategy Funnel/index.csv and existing candidate folders
WHAT YOU ARE PRODUCING TODAY
A minimum of [N] fresh, non-duplicate, due-diligenced candidates. Each gets a
folder under [PROJECT]/Strategy Funnel/candidates/ containing review notes,
source notes, hypothesis, description, code, and implementation notes.
QUALITY BAR
Hard reject: incoherent hypothesis, exact-hash duplicate, no entry/exit rule,
infrastructure unavailable.
Soft warning: out of universe but portable, regime-dependent, heavily
overlapping with existing book.
POST-FLIGHT VALIDATION
ANY miss = redraft twice, then move to rejected/ with a reason.
Required-presence markers: [list every element that must appear in a valid
output file].
Dynamic checks: [list logic-level checks beyond simple presence].
PHASE 2
After Phase 1 completes, run the backtester and optimiser on any candidates
that need it. If the runtime is unreachable or blocked, log phase2_skipped with
a reason and continue. Never let Phase 2 failure block Phase 1 output.
SUCCESS CRITERIA
Phase 1: [N]+ queued candidates; every queued file passes validator checks;
index.csv, queue.md, runs.log, and discovered_sources.md updated.
Phase 2: backtester run or graceful failure logged; optimiser run or graceful
failure logged; dashboard refreshed.
Go.A warning worth heeding
The failure mode: an automated run will take the cheapest path that technically satisfies the instruction unless you close that door explicitly.
- Define the full output scaffold before you automate the task.
- Make the validator stricter than the summary report.
- Reject incomplete output instead of letting it enter the review queue.
- Keep the prompt generic, but keep your reference files specific.
Personal commentary and education, not investment advice or a solicitation. This workflow is provided as-is, with no warranty. Validate all outputs manually before using them in research or live systems. Capital at risk. Past performance is not indicative of future results.