An MCP server that provides on-demand skill loading for AI coding assistants.
BY modbenderUPDATED Oct 28, 2026
INTENDED FOR◈OpenClawInstall in a listed agent.
instructions.md
---
name: open-claw-questrade-426d0487fa4a
description: Execute and monitor stock trades for openclaw.ai workflows using Questrade's browser platform with Yahoo Finance cross-checks. Use when Codex is asked to prepare or review Questrade Web order tickets, monitor symbols from Questrade and Yahoo Finance, produce pre-trade or post-trade checklists, reconcile fills, or enforce privacy-safe trading operations where credentials and sensitive account data stay on the user's side only.
---
# Open Claw
## Overview
Use this skill to run a repeatable browser-trading workflow for Questrade with independent quote checks from Yahoo Finance. Keep execution manual in the broker UI, but automate preparation, validation, and monitoring artifacts.
## Workflow Decision Tree
1. Need a quote/monitoring snapshot across symbols: run `scripts/market_snapshot.py`.
2. Need a trade-ready plan with risk controls: run `scripts/build_trade_checklist.py` with Special Safety Check flags.
3. Need to place a trade in Questrade Web: follow `references/questrade-browser-playbook.md`.
4. Need field-level input/output rules: load `references/data-contracts.md`.
5. Need policy gating details: load `references/openclaw-policy-compliance.md`.
## Standard Execution Flow
1. Confirm trading intent: symbol, side, size, account, session constraints.
2. Pull live Yahoo quotes and optional Questrade export using `scripts/market_snapshot.py`.
3. Reject stale data if snapshot age exceeds user threshold.
4. Run Special Safety Check (mandatory) through `scripts/build_trade_checklist.py`:
- `--policy-ack OPENCLAW_POLICY_ACK`
- `--confirm-user-authorized`
- `--confirm-manual-execution`
- `--confirm-no-secrets-shared`
- Live mode also enforces risk cap, data freshness, and drift thresholds.
5. Draft or refine ticket with `scripts/build_trade_checklist.py`.
6. Enforce hard checks before submitting in browser:
- Confirm side and quantity.
- Confirm order type and time-in-force.
- Confirm max risk and stop/exit rule.
- Confirm buying power and open-order conflicts.
7. Submit manually in Questrade Web and record confirmation ID.
8. Capture post-trade state (fill price, remaining quantity, stop/target status).
## Guardrails
- Never claim an order is submitted unless a human confirms broker submission.
- Treat Yahoo as secondary market data, not authoritative execution data.
- Escalate if broker and Yahoo prices diverge beyond the configured tolerance.
- Refuse ambiguous instructions (missing side, quantity, or symbol).
- Require explicit user acknowledgement before high-impact actions.
- Block checklist generation when safety gate requirements fail.
## Privacy Rules (OpenClaw.ai)
- Keep all credentials, MFA tokens, session cookies, API keys, and passwords user-side only.
- Never ask the user to paste secrets into chat, files, or logs.
- Use masked identifiers in generated artifacts by default (account IDs, order IDs, personal details).
- Only include raw sensitive identifiers when the user explicitly asks and confirms local-only usage.
- Store outputs locally and avoid sharing raw broker exports unless redacted.
## Outputs
Generate one or more of these artifacts per task:
- `snapshot.json` or `snapshot.csv` from `scripts/market_snapshot.py`.
- `trade_checklist.md` from `scripts/build_trade_checklist.py`.
- A concise execution log containing timestamp, symbol, side, quantity, order type, and broker confirmation reference.
## Resource Loading Guide
- Load `references/data-contracts.md` when parsing or validating CSV and JSON payloads.
- Load `references/questrade-browser-playbook.md` when the user asks for browser execution steps, troubleshooting, or reconciliation.
- Load `references/openclaw-policy-compliance.md` when enforcing or explaining policy checks.
- Prefer scripts for deterministic output instead of recreating tables/checklists manually.
---
name: open-claw-questrade-426d0487fa4a
description: Execute and monitor stock trades for openclaw.ai workflows using Questrade's browser platform with Yahoo Finance cross-checks. Use when Codex is asked to prepare or review Questrade Web order tickets, monitor symbols from Questrade and Yahoo Finance, produce pre-trade or post-trade checklists, reconcile fills, or enforce privacy-safe trading operations where credentials and sensitive account data stay on the user's side only.
---
# Open Claw
## Overview
Use this skill to run a repeatable browser-trading workflow for Questrade with independent quote checks from Yahoo Finance. Keep execution manual in the broker UI, but automate preparation, validation, and monitoring artifacts.
## Workflow Decision Tree
1. Need a quote/monitoring snapshot across symbols: run `scripts/market_snapshot.py`.
2. Need a trade-ready plan with risk controls: run `scripts/build_trade_checklist.py` with Special Safety Check flags.
3. Need to place a trade in Questrade Web: follow `references/questrade-browser-playbook.md`.
4. Need field-level input/output rules: load `references/data-contracts.md`.
5. Need policy gating details: load `references/openclaw-policy-compliance.md`.
## Standard Execution Flow
1. Confirm trading intent: symbol, side, size, account, session constraints.
2. Pull live Yahoo quotes and optional Questrade export using `scripts/market_snapshot.py`.
3. Reject stale data if snapshot age exceeds user threshold.
4. Run Special Safety Check (mandatory) through `scripts/build_trade_checklist.py`:
- `--policy-ack OPENCLAW_POLICY_ACK`
- `--confirm-user-authorized`
- `--confirm-manual-execution`
- `--confirm-no-secrets-shared`
- Live mode also enforces risk cap, data freshness, and drift thresholds.
5. Draft or refine ticket with `scripts/build_trade_checklist.py`.
6. Enforce hard checks before submitting in browser:
- Confirm side and quantity.
- Confirm order type and time-in-force.
- Confirm max risk and stop/exit rule.
- Confirm buying power and open-order conflicts.
7. Submit manually in Questrade Web and record confirmation ID.
8. Capture post-trade state (fill price, remaining quantity, stop/target status).
## Guardrails
- Never claim an order is submitted unless a human confirms broker submission.
- Treat Yahoo as secondary market data, not authoritative execution data.
- Escalate if broker and Yahoo prices diverge beyond the configured tolerance.
- Refuse ambiguous instructions (missing side, quantity, or symbol).
- Require explicit user acknowledgement before high-impact actions.
- Block checklist generation when safety gate requirements fail.
## Privacy Rules (OpenClaw.ai)
- Keep all credentials, MFA tokens, session cookies, API keys, and passwords user-side only.
- Never ask the user to paste secrets into chat, files, or logs.
- Use masked identifiers in generated artifacts by default (account IDs, order IDs, personal details).
- Only include raw sensitive identifiers when the user explicitly asks and confirms local-only usage.
- Store outputs locally and avoid sharing raw broker exports unless redacted.
## Outputs
Generate one or more of these artifacts per task:
- `snapshot.json` or `snapshot.csv` from `scripts/market_snapshot.py`.
- `trade_checklist.md` from `scripts/build_trade_checklist.py`.
- A concise execution log containing timestamp, symbol, side, quantity, order type, and broker confirmation reference.
## Resource Loading Guide
- Load `references/data-contracts.md` when parsing or validating CSV and JSON payloads.
- Load `references/questrade-browser-playbook.md` when the user asks for browser execution steps, troubleshooting, or reconciliation.
- Load `references/openclaw-policy-compliance.md` when enforcing or explaining policy checks.
- Prefer scripts for deterministic output instead of recreating tables/checklists manually.
COMPATIBILITY & REFERENCES
Platform guidance
Imported from the listed AgentSkill.sh category. Complete supplied package files and notices are retained. Includes source scripts; review them before use. Required external tools and credentials are not installed. No bundled code was executed during import. Individual agent task behavior has not been tested.
Help users make better hiring decisions. Use when someone is evaluating job candidates, making hiring decisions, conducting reference checks, reviewing work samples or take-homes, calibrating their hiring bar, or deciding between finalists.
Help users navigate career changes and pivots. Use when someone is considering a new role, transitioning into product management, evaluating job offers, taking a sabbatical, or feeling stuck in their current position.
Help users build relationships with mentors and sponsors for career growth. Use when someone is looking for career guidance, wants to find a mentor, needs an advocate at work, is trying to build their professional network, or asking how to get advice from senior leaders.