ib-trailing-stop
Skill411 starsChanged 8 days agoMITView on GitHub
Server-side trailing stop management for stocks and naked LEAPS in IB. Places native TRAIL orders that auto-ratchet the stop as price climbs. Dry-run by default. Requires TWS or IB Gateway running locally.
- Reads credentials
In this repository30 other files it gives its agents.
- CLAUDE.mdCLAUDE.md
- calendarSkill
- chart-ema-rsiSkill
- earnings-calendarSkill
- fundamentalsSkill
- greeksSkill
- ib-accountSkill
- ib-collarSkill
- ib-create-consolidated-reportSkill
- ib-find-short-rollSkill
- ib-option-chainSkill
- ib-pmcc-advisorSkill
- ib-portfolio-action-reportSkill
- ib-portfolioSkill
- ib-report-delta-adjusted-notional-exposureSkill
- ib-stop-lossSkill
- ib-trades-historySkill
- insider-tradingSkill
- markdown-to-pdfSkill
- news-sentimentSkill
- option-chainSkill
- price-historySkill
- report-stockSkill
- risk-assessmentSkill
- scanner-bullishSkill
- scanner-pmccSkill
- spread-analysisSkill
- stock-quoteSkill
- technical-analysisSkill
- whale-huntingSkill
The file
We split this skill into the rules it gives an agent
6 rules. About 0 are common, 3 are distinctive, and 3 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈3
Distinctive · estimate
This author's own choices. Worth a look.
3
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 24
Port fallback: If the configured port fails, automatically retry on the other port.
- Distinctive: line 25
If the retry succeeds, save to memory which account type worked (live/paper) and reuse it for all IB skill calls in this and future …
- Distinctive: line 26
If both ports fail, ask the user to verify that TWS or IB Gateway is running with API access enabled.
- Very specific: line 53
If orphan_orders is non-empty, warn that these were cancelled (execute mode) or need manual cancellation (dry-run).
- Very specific: line 80
Call out preserve_existing separately so user knows nothing changed.
- Very specific: line 148
All analytics live in src/trading_skills/broker/trailing_stop.py:
≈ “Common” is an estimate: at least 1 in 5 of the 364,980 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
IB Trailing Stop Manager
Places IB native TRAIL orders against stocks and naked LEAPS in the portfolio. IB auto-adjusts the stop trigger as the price climbs and locks it as price falls.
PMCC positions are intentionally excluded — use ib-stop-loss for those (a standalone
trailing stop on the PMCC long leg would break the hedge at trigger).
Default mode is dry-run — no orders are placed unless --execute is in the request.
IB Connection
TWS or IB Gateway must be running locally with API enabled:
- Paper trading — port 7497
- Live trading — port 7496
IB_PORTenv var — default port when--portis omitted (e.g.IB_PORT=4001for a Gateway container). Precedence:--portflag >IB_PORT> built-in default. Set it in the shell or a.envfile.
Port fallback: If the configured port fails, automatically retry on the other port. If the retry succeeds, save to memory which account type worked (live/paper) and reuse it for all IB skill calls in this and future sessions — until the user explicitly asks for the other account. If both ports fail, ask the user to verify that TWS or IB Gateway is running with API access enabled.
Instructions
Step 1: Run the script
Dry-run (default — no orders placed):
uv run python .claude/skills/ib-trailing-stop/scripts/trailing_stop.py --symbols JOBY --trail-pct 20Execute (cancel orphan TS_ orders + place new TS_ TRAIL orders):
uv run python .claude/skills/ib-trailing-stop/scripts/trailing_stop.py --symbols JOBY --trail-pct 20 --executeExecute forced (cancel + replace existing TS_ orders with current parameters):
uv run python .claude/skills/ib-trailing-stop/scripts/trailing_stop.py --execute --forcedStep 2: Format the report
Format JSON output as a markdown report with three sections:
Section 1: Existing TRAIL Orders
Show all_trail_orders.module (TS_ orders) and all_trail_orders.manual (manually placed).
If orphan_orders is non-empty, warn that these were cancelled (execute mode) or need manual cancellation (dry-run).
Section 2: Positions
For each entry in positions, show a table:
| Field | Value |
| Symbol | JOBY — stock (1000 shares) |
| Spot | $7.50 |
| Reference | $7.50 (max of current $7.50, avg cost $5.00) |
| Trail params | 20% (initial stop $6.00) → action: place_new |
| Existing trail | none |
For LEAPS rows, also show the option leg (strike/expiry) and current option mark.
action values:
place_new— no existing TS_ order; one will be created in execute modepreserve_existing— TS_ already in place; left alone (IB has been tracking the high)overwrite—--forcedis on; existing TS_ will be cancelled and replaced
Section 3: Order Results (execute mode only)
Per-position result with order_id and order_ref.
Step 3: Report to user
- State dry-run vs execute mode prominently.
- Lead with positions that have
action: place_new(these are about to get a new TRAIL). - Call out
preserve_existingseparately so user knows nothing changed. - Show existing trail details (current
trail_stop_price) so user knows where IB has trailed to.
Arguments
| Flag | Default | Description |
--port | 7497 | IB Gateway/TWS port |
--account | all | Specific account ID |
--symbols | all | Analyze only these symbols |
--trail-pct | 20 | Trail amount as percentage of reference |
--trail-amt | — | Trail amount in dollars (mutually exclusive with --trail-pct) |
--price-mode | mid | Option pricing: mid or last (LEAPS only) |
--execute | off | Cancel orphans + place TS_ TRAIL orders |
--forced | off | Cancel and replace existing TS_ orders (requires --execute) |
JSON Output Structure
{
"generated_at": "2026-05-29 10:00 ET",
"data_delay": "real-time",
"dry_run": true,
"forced": false,
"trail_pct": 20.0,
"trail_amt": null,
"accounts": ["U1234567"],
"symbols_filter": ["JOBY"],
"all_trail_orders": {"module": [], "manual": []},
"orphan_orders": [],
"positions": [
{
"symbol": "JOBY",
"type": "stock",
"account": "U1234567",
"qty": 1000,
"underlying_price": 7.50,
"stock": {
"avg_cost": 5.00,
"current_price": 7.50
},
"trail_stop": {
"trail_pct": 20.0,
"trail_amt": null,
"reference": 7.50,
"initial_stop_price": 6.00,
"action": "place_new",
"existing_trail": null
}
}
]
}Key Behaviors
- Reference price =
max(current_price, avg_cost)— locks in profit when above cost, never starts below entry. - Forced mode uses
current_priceas the reference (can place an initial stop below entry, useful when re-arming after a drawdown). - Existing TS_ orders are preserved by default because IB has been ratcheting the trail since placement — replacing would reset that tracked high. Use
--forcedto deliberately reset. - Scope: stocks + naked LEAPS only. PMCC positions are excluded; use
ib-stop-lossfor those.
Order Identification
TS_{SYM}_{STRIKE}_{EXPIRY}_{RIGHT}— naked LEAPS TRAIL orders (right isCorPso calls and puts on the same strike/expiry don't collide)TS_{SYM}_STK— stock TRAIL orders
Architecture
All analytics live in src/trading_skills/broker/trailing_stop.py:
Analytics (no IBKR — testable in isolation):
calc_trail_reference—max(current_price, avg_cost)normally;current_priceif forcedcalc_initial_trail_stop_price— reference × (1 − trail_pct/100) or reference − trail_amtidentify_trailable_positions— stocks + naked LEAPS; PMCC excludedbuild_trail_analysis— full per-position output dictdetect_orphan_trail_orders— TS_ TRAIL orders for gone positionssummarize_all_trail_orders— splits IB TRAIL orders into module vs manual
Data layer (IBKR):
get_trailing_stop_data— main entry point_cancel_orphan_orders— cancel stale TS_ orders_place_simple_trail_order— native TRAIL order on stock or option_execute_position_trail— dispatch per position type
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.

