Skip to content

aaditya-v-more/worldquant-orchestrator / skills/wqo-cli

wqo-cli

Flagged

Set up and use the WQO command-line client for WorldQuant BRAIN research. Use for CLI installation, authentication onboarding, command discovery, or research workflows spanning data discovery, simulations, analysis and explicitly confirmed submission.

FileKindSize
SKILL.mdskill md2.3 KB
references/runtime.mdmarkdown3.0 KB
scripts/wqo.shexecutablescript2.0 KB

references/runtime.md

View on GitHub
# Setup and operational boundaries

Use macOS or Linux. Run from the user's working project, not the skill source
repository. Resolve `scripts/wqo.sh` relative to the installed SKILL.md and invoke
it by absolute path with Bash. Commands written as `wqo ...` in the skill mean
`bash /absolute/path/to/this-skill/scripts/wqo.sh ...`.

The runner reuses a compatible WQO CLI on PATH or in the project's `.wqo/venv`.
When missing, it installs `wqo==0.1.1` from PyPI into that project environment,
using uv or Python 3.12+. It never installs skills or the CLI globally, edits
shell profiles, or changes credentials. If both prerequisites are missing, show
the prerequisite error; do not invent an installer. `WQO_PROJECT_DIR` can select
an existing project explicitly. Keep `.wqo/` and installed agent skill folders
out of Git. The runner preserves CLI arguments, JSON output and exit codes.

Before the first live operation, run `wqo state` and `wqo auth status`. These
skills do not assume a particular account, level, operator count or concurrency.
Use the installed CLI's `--help` for flags. `account snapshot` requires WQO 0.1.1+;
use it only when `account --help` lists it.
Otherwise use `auth status` and `account status`. Raw `api` accepts only reads.
Expressions and settings are sent to BRAIN for simulation. Example expressions
are research hypotheses, and backtest results do not guarantee future returns.

- Never submit without an explicit yes in this conversation for that specific
  alpha. Installation, mining and a passing report do not authorize submission.
- Never read back, write, echo or log `~/.brain_credentials.json`. If absent,
  give the user the JSON format `{"email":"...","password":"..."}` and have
  them create it privately in a text editor, then `chmod 600` it. Only WQO's
  normal authentication flow uses the credentials.
- On Persona verification, give the inquiry URL to the user and wait until they
  confirm completion before `wqo auth persona`. Never automate or bypass it.
- Exit 3 means stop at the quota. Never raise budgets/concurrency or weaken gate
  thresholds to work around a failure. Unknown submissions retain quota across
  days; inspect the outcome instead of retrying or deleting the ledger.
- Reuse the existing per-user ledger (`wqo state`). When upgrading a checkout,
  stop other WQO processes and use `wqo state --migrate-from /old/checkout/data`
  or keep `WQO_DATA_DIR` pointed at that ledger. Never start a new ledger to evade
  quota, and do not change its location just because skills were installed.
- Before a BRAIN metadata write, check names/tags/descriptions for tool labels,
  mining template IDs or test names. Keep provenance local, not in BRAIN fields.
- Keep real alpha expressions, results and account details out of tracked files
  and public posts. Report failed candidates and unavailable checks honestly.

Exit codes: 0 success; 1 error/gate blocked; 2 auth/Persona; 3 budget exhausted;
4 submission refused. Inspect these statuses rather than assuming any JSON means
success. CLI setup alone does not spend simulation or submission quota.
add-product-e2e-evalpaperclipai/paperclipAdd or extend a Paperclip full-stack runner E2E workflow, fixture, matcher, or report evidence path for local or Daytona execution.Stars: 115kUpdated Claude CodeCodexVerifiedInstall
add-runner-evalpaperclipai/paperclipAdd or extend a Paperclip Runner protocol evaluation definition, roster, assertion, or report fixture with provenance and narrow validation.Stars: 115kUpdated Claude CodeCodexVerifiedInstall
diagnose-why-work-stoppedpaperclipai/paperclipCompatibility entry point for failed, stalled, looping, or over-recovered Paperclip work. Use fix-failed-paperclip-run for the full diagnosis, regression, repair, real verification, and green-PR workflow; honor an explicitly requested diagnosis-only scope.Stars: 115kUpdated Claude CodeCodexVerifiedInstall
fix-failed-paperclip-runpaperclipai/paperclipDiagnose and fix failed, stuck, slow, or incorrectly blocked Paperclip tasks and runs. Use with a task/run URL or failure report to trace the root cause, prove a regression red then green, check second-order effects, verify the real workflow, and deliver a PR with passing checks.Stars: 115kUpdated Claude CodeCodexVerifiedInstall