Best FTMO EA Review: Automated Challenge Passing

What “best” means for an FTMO challenge EA

A useful review begins by refusing a universal ranking. One EA may suit a trader who can observe London-session entries every day, while another suits a swing trader who needs fewer interventions. For FTMO, best means compatible with the exact programme, platform, instruments, and risk constraints you intend to use. It does not mean the highest profit factor on a vendor landing page. The candidate must make decisions using prices and liquidity that the account can realistically receive, and it must have a loss pattern that leaves room for normal execution error.

Score each candidate across five questions. Can you explain why it enters and exits? Is the result history long enough to include different conditions? Does its maximum adverse period fit a deliberately reduced risk setting? Can it enforce a stop even if you are asleep or disconnected? Can you operate it without prohibited coordination, shared access, or attempts to exploit infrastructure? A robot that fails one of these questions is not rescued by a polished equity curve. This is especially true where a system relies on rapid recovery after loss, hidden averaging, or a rare but ruinous basket.

Separate a strategy review from a sales review. Sales material tells you what a vendor wants you to notice. A strategy review asks what happens after three losses, a server restart, a spread spike, a holiday session, and a parameter change. Record answers in a comparison sheet before opening an account. For selection principles that apply beyond one firm, see how to compare challenge EAs.

• Write the exact FTMO programme and platform you are reviewing.
• Reject any claim of guaranteed passing, guaranteed payout, or zero risk.
• List the EA’s instruments, holding style, maximum simultaneous trades, and intended session.
• Mark every feature you cannot explain as a research item, not an advantage.

Start with current official eligibility and terms

An EA can be technically excellent and still be unsuitable if the account holder, country, payment route, or strategy use is ineligible. Before paying, check FTMO’s official current restrictions for your residence, identity verification, payment method, platform access, and the programme selected. Do not rely on a video, affiliate review, forum comment, or an old screenshot. They may describe a former rule or a different account type. Keep the confirmation email, invoice, and a dated copy of the terms in an account folder.

Automation permission is not a blanket approval for every automated behavior. Review the provisions concerning Expert Advisors, prohibited trading practices, account ownership, third-party assistance, copying, technical exploitation, and any funded-stage limitations. If your EA consumes an external signal, uses a trade copier, trades around news, or shares a VPS with another trader, describe that specific workflow to support and ask whether it is permitted. A vague question such as “are bots allowed?” cannot settle a detailed compliance issue.

Eligibility also has a practical financial side. Confirm the currency and method used for the evaluation purchase, possible card or conversion charges, and the documentation required for a future payout. Payout receipt can create tax, reporting, business-registration, or banking obligations in your own jurisdiction. FTMO cannot replace local legal, tax, or financial advice. Maintain trade statements, invoices, payout records, and exchange-rate records from the start rather than trying to reconstruct them after a payment.

Read performance evidence like an operator

Ask for a full report rather than a cropped balance chart. Relevant evidence identifies whether results are backtested, demo-forward-tested, evaluated, or funded; shows the date range; reports trade count; and reveals open-trade drawdown as well as closed results. It also identifies symbols, timeframe, broker conditions, commissions, swaps, and settings. A short run during a favorable trend says little about resilience. A review should be able to locate the largest losing sequence and describe how the EA behaved without changing the story.

Inspect concentration. If a few trades create most gains, the expected path to a target may be irregular and difficult to manage. If losses are small but the occasional loss is enormous, the system may be averaging or deferring risk. If returns leap during a particular news release or rollover hour, execution assumptions may be doing more work than the strategy. Compare maximum equity drawdown, not merely balance drawdown, because floating loss can matter operationally even before a position closes.

Forward evidence on a similar platform is more valuable than an optimized historical report on an unrelated feed. It need not be perfect or profitable every month, but it should be auditable and honest about deviations. Ask whether trades were manually removed, whether settings changed mid-series, and whether the reported account used materially different leverage or symbols. The companion backtesting versus live EA trading article explains why an attractive test is a hypothesis, not proof.

• Request or export a report with closed and floating drawdown information.
• Identify the longest losing run and largest single loss.
• Compare spread, commission, symbol suffix, and session assumptions with your intended platform.
• Label every result by source: test, demo, evaluation, or funded account.

Understand the strategy before trusting automation

You do not need source code to understand a strategy’s economic behavior. A vendor should state whether the EA is trend-following, mean-reverting, breakout-based, grid-based, arbitrage-sensitive, or a mixture. Ask what invalidates an entry, what closes a losing trade, whether a hard stop is sent to the server, and whether additional positions are used to repair a basket. Plain answers permit risk analysis. Evasive phrases such as “AI precision” or “secret institutional logic” do not.

Trend systems can accept many small losses before a large move, so their challenge risk lies in the losing streak and patience required. Mean reversion may look smooth until a sustained directional move exposes its tail risk. Breakout systems can be sensitive to spread and false breaks. Grid and martingale-like recovery logic can concentrate risk precisely when volatility rises. None is automatically good or bad, but each needs a setting that matches its failure mode rather than its best month.

Look for behavior that remains understandable when conditions are unpleasant. A robust EA can state its maximum number of entries, its maximum aggregate exposure, and the condition that makes it stand down. It should not need ever-increasing volume, an undefined emergency close, or manual rescue to preserve its record. For a broader map of suitable system families, use this algorithmic strategy guide.

Translate FTMO limits into an internal risk budget

Use official limits as outer legal boundaries, not targets for the robot. Confirm the current loss definitions, reset time, and whether calculations include floating profit and loss, fees, commissions, or swaps for your programme. Then build an internal daily stop and account stop materially inside those boundaries. The size of the buffer is a business decision based on volatility, expected slippage, and correlated positions, not a number copied from a stranger’s set file.

For example, imagine an EA can open three related currency positions after one signal. The apparent risk per order is misleading because the positions can lose together. Convert every open and pending trade into a worst-case account-level loss at its stop, add plausible spread expansion and execution delay, then compare the total with your internal stop. If the basket could consume most of the day’s room, reduce volume, reduce concurrent trades, or restrict the symbols. Do not assume a diversification benefit where currencies share the same macro driver.

A proper configuration has several brakes: per-trade stop loss, maximum lots, maximum positions, symbol exposure cap, daily realized-and-floating loss stop, overall equity stop, and a manual emergency procedure. Verify that the EA’s interpretation of account currency and contract size is correct. The EA drawdown and lot-size guide is useful for checking assumptions, but the official account dashboard and terms remain authoritative.

• Set internal stops below, never at, the published boundaries.
• Include floating loss, costs, correlated trades, and slippage in the worst-case estimate.
• Test that each stop disables new entries as well as closing or managing old ones.
• Document who can manually disable the EA and how.

Review position sizing and recovery behavior

Position sizing tells you more about challenge suitability than a headline win rate. A modest win rate can be workable with defined losses and stable sizing. A high win rate can be dangerous when one recovery basket erases months of gains. Open the inputs and search for multiplier, recovery, grid distance, maximum levels, equity protection, and lot cap. If the developer cannot state the largest possible exposure under ordinary settings, you cannot responsibly match it to an evaluation.

Fixed lots are not automatically safer than percentage sizing. A fixed lot may become too large on a smaller account or a different symbol. Percentage sizing can become too aggressive after gains if the daily-loss framework uses a reference you have not modeled. The safer choice is the one that makes maximum loss per idea predictable after contract specifications, stop distance, and simultaneous exposure are considered. Test the intended account size rather than extrapolating from a smaller demo.

Recovery after a loss deserves a separate decision. Some systems increase size, add entries, or delay the exit in pursuit of a mean reversion. That can transform a routine losing run into a sudden breach. Prefer an EA that can continue at reduced risk or pause after a defined sequence. A challenge does not reward emotionally satisfying recovery; it rewards surviving the next eligible trading opportunity.

News, spreads, rollover, and market closures

A credible FTMO EA review treats market microstructure as part of the strategy. An entry logic that is valid at a normal spread may be invalid during a major release, session opening, holiday-thinned liquidity, or daily rollover. Check current official policy for the programme and stage before assuming that news trading or holding through an event is allowed. Even where permitted, the practical question remains whether the robot was tested under comparable spread and slippage conditions.

Configure a calendar filter from a reliable source and verify its clock. The filter should specify currency relevance, impact classification, blackout duration before and after, and its behavior if the calendar cannot refresh. A safe failure mode is to block new entries, notify you, and continue managing existing protective orders according to the written plan. Do not turn off a filter simply because it missed a profitable event in a backtest.

Rollover needs equal care. Swaps, wider spreads, reduced liquidity, and platform maintenance can make a short-term strategy trade differently around the broker’s daily boundary. Test the EA’s behavior during closure and reopening, including pending orders and trailing stops. Read the guide to gaps, slippage, and spread before allowing an unattended system to hold risk through a weekend or holiday.

UTC, server time, and the hidden clock problem

Every time-based EA has at least two clocks: the computer or VPS clock and the trading server clock. Economic calendars often use UTC or local release time, while a platform may display broker-server time that changes with daylight-saving conventions. A session filter set for “08:00” is meaningless until you establish which clock it uses. Record the relationship among UTC, server time, and your local time on the launch date, then review it when daylight-saving changes occur.

The daily-loss reset is another reason not to improvise with local midnight. Confirm the current official reset convention and observe the account dashboard rather than estimating. Schedule the EA’s daily reset, profit lock, and reporting window with a margin around the server boundary. A position opened near reset can span two accounting periods, and floating loss can make a seemingly harmless hold materially different from a closed-trade journal entry.

Perform a clock drill on demo. Force a session opening, news blackout, daily reset, and end-of-week shutdown using the exact terminal and VPS. Capture timestamps in screenshots and journal notes. If the EA has no transparent way to state its time basis, that uncertainty is a material defect. Time alignment is also discussed in the FTMO EA setup guide.

• Write down UTC, platform-server, VPS, and local time at setup.
• Check daylight-saving transitions for every timed rule.
• Confirm the programme’s current daily reset from official material.
• Test calendar failure, terminal restart, and session shutdown behavior.

Platform, symbol, and execution compatibility

An EA that works on one MetaTrader installation may fail subtly on another. Verify the supported platform, build, account login type, symbol names, contract specifications, minimum and maximum volume, stop-distance rules, and permitted order types. Symbol suffixes, five-digit pricing, metals contract size, and index sessions can alter an algorithm’s calculations. Never assume that a familiar chart label represents identical conditions across brokers or account servers.

Install only a legitimate, licensed copy obtained from a traceable source. Scan files according to your security practice and review what the EA requires: DLL imports, web requests, licence server access, external files, or copy-trading bridges. A requirement for unexplained remote control is a serious security and compliance concern. Change platform passwords through official channels, enable available account security measures, and do not provide credentials to a vendor who offers to operate the account.

Execution testing should compare the intended settings with real quotes. Watch rejected orders, requotes, invalid stops, partial behavior, lot rounding, and duplicate entries after reconnection. A small number of test trades is not proof of edge, but it can expose installation errors before they become an evaluation loss. Platform choice details are covered in the MT4 versus MT5 EA comparison.

VPS reliability and active monitoring

A VPS is useful when it improves availability, not when it creates the illusion that supervision is unnecessary. Select a location and specification appropriate to the platform and obtain it from a reputable provider. Keep the operating system, terminal, and EA updated deliberately, not in the middle of an active trade. Disable sleep, confirm automatic restart behavior, and ensure the terminal can reconnect without opening a second instance of the strategy.

Create alerts for terminal disconnection, account authorization failure, disabled automated trading, abnormal drawdown, missing ticks, and unexpected order count. Check the Experts and Journal tabs after any interruption. A notification should lead to an action plan: identify whether orders remain protected on the server, compare actual positions with the EA’s state, and decide whether to disable new trading. Do not blindly restart a robot that may interpret existing positions as fresh signals.

Use a simple daily operating ritual: inspect account status before the relevant session, check the event calendar and server time, verify inputs and exposure, then review the order log afterward. Keep a local backup of templates and set files, but do not share account access. For practical hosting considerations, see the current forex EA VPS guide.

• Test reboot, reconnect, and terminal update behavior before launch.
• Enable alerts for disconnection, drawdown, and unexpected trades.
• Back up the approved set file with its date and account context.
• Keep credentials private and limit remote access to the account holder.

A fair demo and forward-test protocol

Forward testing should answer an operational question: does this exact EA, on this platform, with these inputs, behave close enough to the reviewed evidence to justify an evaluation? Use a practice environment that resembles the intended account where possible. Match symbols, leverage assumptions, trade permissions, risk setting, session schedule, and news filter. Do not optimize settings after every losing day. That practice fits the sample and destroys the information the test was supposed to provide.

Define pass criteria before testing. Examples include no unexplainable order behavior, hard stops functioning, exposure staying within the internal budget, calendar pauses occurring at the right time, and drawdown remaining consistent with the model. Define stop criteria too: a missing protective stop, an unexplained duplicate trade, a deviation caused by a platform mismatch, or a loss sequence beyond the tested profile. A stop criterion protects money and provides a concrete issue for the vendor or your own investigation.

Journal every intervention. Note whether a trade was closed manually, whether a filter was overridden, and whether the terminal was offline. Those entries prevent you from attributing human discretion to the EA. A clean, uneventful test is valuable because it verifies the plumbing; it does not convert probability into certainty. When it is complete, compare the findings with the original evidence rather than with your hope of starting quickly.

Launch an evaluation without chasing the target

The first evaluation week is not the time to turn up risk because a target looks distant. Start with the same approved configuration used in the final test, or lower exposure further while you confirm the live environment. The account fee is already spent; increasing risk to recover it only makes the next decision worse. A small early gain does not prove that the EA is safe, and a small early loss does not prove that it has failed.

Set a cadence for review instead of watching every tick. At a suitable daily time, inspect open exposure, internal drawdown, trade log, scheduled events, and connection state. At the end of the week, compare realized behavior with the plan: trade frequency, loss size, session adherence, and any manual action. If a material mismatch appears, pause and diagnose before resuming. Changing lot size, filters, and entry logic simultaneously makes diagnosis impossible.

Treat the profit target as an outcome, not a command. An EA with a realistic edge may need time, and periods with no trades can be entirely correct. Avoid adding unrelated robots just to create activity. Multiple systems can share exposure, compete for margin, or produce matching patterns that warrant compliance questions. The disciplined alternative is a single documented system with a known risk budget.

Account access, shared signals, and compliance hygiene

The account holder should retain control of credentials, payment records, and trading decisions. Do not use a passing service, buy access to another person’s account, or allow a vendor to log in and operate yours. Even an honest-looking arrangement can conflict with account ownership or third-party access provisions. If technical support needs to diagnose an EA, provide logs or a screen recording with sensitive details removed, or use a separate test environment rather than handing over the live login.

Identical or highly synchronized orders across accounts can raise questions, particularly where a common signal seller, copier, or shared configuration is involved. That does not mean a legitimate user must invent different behavior, but it does mean you should understand the current policy and avoid arrangements intended to disguise coordination. Maintain a record of your EA licence, settings, VPS, devices, and support correspondence. If FTMO requests information, respond accurately and promptly rather than editing history.

Travel and a VPS can be legitimate, but they change the access story. Tell support before a material access change if official guidance requests it, and confirm country eligibility before working from a new location. The detailed FTMO IP and VPS guide explains the operational questions to settle. Compliance is not an obstacle to outsmart; it is a condition of being paid.

Cost, licence, payment, and payout realism

Compare total operating cost rather than a sticker price. Include the evaluation fee, resets if relevant, EA licence or subscription, VPS, optional data service, currency conversion, card fees, and time spent monitoring. A cheap robot that needs frequent paid parameter updates or a large account to recover from drawdown may be more expensive than a transparent, lower-risk alternative. Ask whether the licence is tied to a machine, account number, expiry date, or recurring payment.

Read cancellation, refund, update, and support terms before paying the EA vendor. A lifetime licence may exclude meaningful upgrades; a subscription may continue charging while you pause trading. Payment through a method with no clear merchant identity or receipt is a warning sign. Keep invoices and correspondence. Do not use payment details, identities, or accounts belonging to another person to bypass eligibility or payment restrictions.

Passing does not make a payout automatic. Current FTMO requirements, trading conditions, identification, payout cycles, profit split, payment options, and any minimums should be confirmed from official sources at the time. Plan cash flow conservatively. A funded account can have a different policy environment from the evaluation, and local tax or reporting duties can arise when funds arrive. Consult a qualified local adviser for your situation.

• Calculate one-year operating cost, including licence and hosting.
• Save vendor and FTMO invoices, terms, and support confirmations.
• Verify payout process and available payment method when it becomes relevant.
• Keep records needed for local tax, banking, or reporting advice.

Warning signs that disqualify an FTMO EA

Disqualify vendors who guarantee an FTMO pass, conceal loss mechanics, pressure immediate purchase, or insist that you disable safety controls. Be skeptical of only-selected screenshots, anonymous testimonials, impossible smoothness, and results with no timestamps or trade list. A vendor does not need to disclose proprietary code, but they should be able to disclose the strategy category, risk controls, supported platforms, required maintenance, and known limitations.

Technical red flags include no hard exit, unlimited averaging, no maximum lot, dependence on a specific unverified broker feed, random changes to settings, and a licence that locks you out during an open trade. Operational red flags include requests for your password, remote access without a clear purpose, pirated software, hidden recurring charges, and instructions to obscure IP or account use. None of those features improves an edge; they increase fragility and compliance risk.

Beware of ratings that rank systems solely by return. A review that ignores equity drawdown, strategy logic, sample length, costs, and rule fit is advertising dressed as analysis. Use the wider EA settings optimization guide to assess settings carefully, then favor the candidate with the clearest downside over the candidate with the most exciting headline.

From passing to sustainable funded operation

The evaluation is only a qualification period. Before moving the EA to a funded account, reread the current official terms for that stage and compare them with the evaluation rules you modeled. News handling, payout conditions, permitted practices, risk calculations, and programme features may not be identical. Rebuild the compliance checklist rather than assuming that the old set file carries every assumption forward.

Funded operation benefits from lower urgency. Set a monthly objective around rule-compliant execution and capital preservation, not an attempt to reproduce a lucky challenge week. Withdrawals, if and when available under current rules, should be planned with awareness of the remaining buffer and your personal cash needs. Do not immediately increase size after a payout or a profitable streak. A stable process produces better information than constant scaling.

Review the system monthly: compare expected and actual drawdown, check vendor updates, revalidate time filters, inspect costs, and archive statements. If the market behavior that supported the edge has changed, pausing is a legitimate decision. The best FTMO EA review ends with this conclusion: the best choice is a transparent, bounded, actively supervised tool that fits verified current terms, not an automated promise of challenge passing.

A funded account should also change the way you evaluate opportunity cost. During an evaluation, a trader may be tempted to pursue every eligible setup because the target is visible. Once funded, preserving a clean operating record and receiving a valid payout can be more valuable than forcing an extra trade into a poor liquidity window. This does not mean that the EA must be permanently timid. It means its active risk setting must reflect the account’s actual current buffer, the quality of market conditions, and the fact that a breach can interrupt a process that was otherwise working. Keep a written distinction between a normal strategic drawdown and a deterioration that requires review. The former calls for adherence; the latter calls for a pause and evidence.

Periodic review should include execution quality, not only profit. Compare requested price with fill price, observed spread with the test assumptions, stop placement with the intended distance, and actual holding time with the design. A strategy can remain logically sound while its net outcome weakens because costs or liquidity have changed. Conversely, one exceptional month with unusually favorable fills should not cause a risk increase. Export statements at a regular interval and annotate material platform incidents, vendor updates, and manual actions. This creates a usable audit trail when you ask whether performance changed because of the market, the broker environment, the configuration, or the operator. It is much stronger than relying on memory or a balance curve.

Keep personal administration separate from the trading terminal. Reconcile evaluation and payout documents, retain receipts for services, and review whether your address, identity, bank details, or tax residence has changed. If you move country, change payment route, travel extensively, or receive a request for verification, consult current official guidance and respond truthfully. Local obligations vary widely, and terminology such as contractor income, trading income, service income, VAT, foreign exchange reporting, or capital treatment can have different meanings in different jurisdictions. A qualified adviser who understands your local facts is the right source for personal reporting advice. Good recordkeeping does not guarantee a particular treatment, but it makes informed advice possible.

Finally, decide in advance what would make you stop using the EA. Examples include a material change in official terms, a vendor update you cannot validate, a persistent execution mismatch, a drawdown beyond the tested range, a security concern, or a strategy behavior you cannot explain. A stop decision is not an admission that automation failed. It is evidence that the trader, rather than the software, remains accountable. Save the final review, schedule the next review date, and retain the ability to disable new entries immediately. That is the durable standard for automated challenge passing: measured evidence before launch, explicit controls during operation, and a calm exit process when assumptions no longer hold.

There is a useful difference between automation and abdication. Automation applies pre-decided rules with speed and consistency. Abdication occurs when a trader cannot describe what the program may do, has no alerting, has not checked current terms, or leaves the only copy of the configuration with a seller. The first can improve discipline; the second creates unmanaged operational risk. Build procedures that preserve the benefits of automation while retaining human accountability. Know the broker-server time, know the current exposure, know how protective orders are placed, and know which controls remain effective when the terminal is offline. Those answers are more valuable than a new indicator or a fashionable label.

When comparing candidate EAs, use the same worksheet and the same skeptical standard for every one. Record evidence period, instrument universe, estimated execution sensitivity, maximum observed equity drawdown, worst modeled basket, stop mechanism, news behavior, licensing, support response, and official rule questions. Then make a decision based on weaknesses as well as strengths. A candidate with fewer unknowns is often the more professional choice even if another candidate displays a larger historical return. This framework is deliberately slower than purchasing from a ranking table, but it prevents selection from being driven by affiliate promotions, recency bias, or an isolated success story.

If your first evaluation does not pass, preserve the records and conduct a factual post-mortem before purchasing another. Identify whether the result came from a normal loss sequence, an implementation error, a rule misunderstanding, excessive risk, a changed market condition, or a combination. Do not erase logs because the result is disappointing. The purpose is to decide whether the original hypothesis remains credible at a lower setting, needs a corrected test, or should be rejected. Repeating the identical unreviewed configuration is not persistence. A measured review, with current terms checked again, is the only defensible basis for another attempt.

Build a vendor due-diligence interview

Before buying, send the vendor a short written questionnaire and judge both the answers and the omissions. Ask which platform builds are supported, whether the EA requires DLL imports or web access, what happens if the licence server is unavailable, which symbols were used in the presented evidence, and whether a hard stop is attached to every entry. Ask for the maximum concurrent exposure under the recommended preset, not merely the nominal risk percentage. A serious vendor may protect proprietary details, but should still answer operational questions precisely.

Ask how updates are distributed and whether an update can alter entry logic, risk settings, or licence validation while you have an open position. Establish whether support covers installation, parameter interpretation, and defect investigation, or only payment access. Confirm the refund terms in writing. An honest answer that says a strategy can experience difficult periods is more useful than a vague reassurance that it has always recovered. Your aim is to establish whether the seller understands the product as a risk system rather than as a marketing funnel.

Use the response to create an acceptance record. Include the vendor name and contact route, version number, purchase date, licence terms, recommended inputs, known limitations, and the evidence links you reviewed. Do not accept an invitation to share your FTMO login so that someone else can configure or trade the account. You can ask for installation guidance while preserving sole control. This record makes future troubleshooting factual and helps distinguish an approved setting from a later improvisation.

• Ask for maximum basket exposure and maximum number of positions in writing.
• Confirm whether every trade has a server-side protective stop.
• Record EA version, preset version, licence duration, and update policy.
• Refuse any vendor request for your live account credentials.

Make the final purchase decision

Make the decision only when the evidence file, compliance check, test protocol, and operating plan agree. The strongest candidate is often the boring one: it has modest and inspectable claims, an understandable stop mechanism, a narrow set of supported conditions, and a vendor who does not require secrecy from FTMO or access to your account. A lower expected return can be superior when its loss behavior is bounded and its implementation is repeatable. Challenge fees should be money you can afford to lose, not capital required for bills.

Use a simple decision rule. Approve only if current official FTMO terms appear compatible, relevant performance evidence is sufficient for your own standard, maximum modeled exposure fits the internal buffer, a matched forward test passes the predefined checks, and you can personally monitor the system. Defer if any answer is uncertain. Reject if the candidate requires prohibited behavior, unexplained recovery risk, misleading evidence, or a vendor relationship that compromises account control. Delay is cheaper than a preventable breach.

Once approved, freeze the launch version. Store the chosen preset, chart and symbol list, calendar settings, time conversion, VPS details, and emergency contacts. Write the date on the plan and note which official pages were checked. Revisit it when FTMO updates terms, the vendor releases a revision, or daylight-saving and market conditions change. That disciplined closure is the practical meaning of a best-FTMO-EA review: a decision you can explain, execute, audit, and stop safely.

• Approve only after rule fit, evidence, risk model, and forward test all agree.
• Keep the evaluation fee separate from essential personal expenses.
• Freeze and archive the approved launch configuration.
• Schedule future checks for official policy, platform, and EA updates.