メニュー
Use when expanding Chaos Redux event ideas into detailed specifications before implementation.
Codex または Claude でインストール この Prompt をコピーして Codex、Claude、または他のアシスタントに貼り付けると、Skill ページを確認してインストールできます。
SOC 職業分類に基づく
Use when designing, implementing, auditing, or fixing Hearts of Iron IV decisions, missions, timed objectives, decision categories, and mission localisation.
Use when designing, implementing, auditing, or fixing Hearts of Iron IV national focus trees.
Use when implementing or updating Chaos Redux events.
Use when creating, sourcing, processing, converting, organizing, wiring, or documenting visual assets for Chaos Redux.
Use when designing, researching, wiring, or documenting Chaos Redux super-events.
Use when coordinating custom Codex subagents for Chaos Redux implementation, asset production, super-event research, audits, active small patches, planning handoffs, or documentation work.
| name | chaos-redux-event-planning |
| description | Use when expanding Chaos Redux event ideas into detailed specifications before implementation. |
Use this skill to design or expand events for the Hearts of Iron IV mod Chaos Redux.
This skill creates event specifications directly inside the Chaos Redux repository. It does not implement code. Implementation belongs to chaos-redux-events. Visual asset generation and processing belongs to chaos-redux-event-assets. Animated sprite planning, frame-sheet requirements, animated portrait packages, and animation handoff details belong to chaos-redux-frame-animation when motion is needed. Super-event quote, remark, music, and presentation research belongs to chaos-redux-super-events.
Before writing the event specification, use the following as the design baseline:
AGENTS.mdchaos-redux-eventschaos-redux-event-assets when the event needs visual assetschaos-redux-frame-animation when the event has animated sprites, animated UI, animated route emblems, animated portraits, warning pulses, hover loops, glow loops, float loops, particle loops, or frame-by-frame presentation needschaos-redux-super-events when the event needs a super-eventhoi4-focus-trees or the current focus-tree skill when the event needs focus treeshoi4-decisions-missions when the event needs decisions, missions, timed objectives, influence actions, or decision-driven mechanicsUse those sources to understand how Chaos Redux already works, then expand creatively from there.
Inspecting the repo is mandatory, but the specification should not read like a technical audit. Use repo findings to prevent mistakes and match patterns. Record proof of inspection in the completion report, not in admin sections inside the spec.
The user will provide either:
Treat the user idea as the starting point. Do not assume it is final.
When the idea is rough, expand it into a deeper design.
When the idea is already detailed, preserve the core direction and improve structure, clarity, missing connections, and implementation readiness.
The goal is to create an event that feels like a layered Chaos Redux system, not a basic popup.
The event should have atmosphere, player choice, consequences, escalation, lore, replay value, and enough visual support to feel finished.
The design should be ambitious. Do not get lazy, conservative, or minimal unless the event is genuinely small. Depth matters more than speed. Take the time needed to research, compare patterns, think through consequences, and design the event as if the coding agent will implement exactly what is written.
The finished spec should make the coding agent feel that the event has already been designed in full.
The spec should focus on the event idea, not on obvious implementation plumbing.
Do not put sections such as:
ScopeSource baselineRepository contextSpreadsheet rowExisting implementation auditGeneric trigger safeguardsThe spec should be player-facing design and implementation-relevant event design.
Avoid obvious lines such as:
Those are baseline system responsibilities. Include technical notes only when they prevent a likely mistake, explain non-obvious behavior, or define a unique rule for this event. Otherwise, its just noise.
Do not create negative capability notes for absent event surfaces. If an event does not have a world-end scenario, do not mention world-end scenarios. If an event does not have a manual triggerable scenario, do not mention triggerable scenarios. Do not write sections, bullets, event-detail notes, spreadsheet-facing summaries, implementation prompts, or player-facing text that say no world-end scenario, does not have a world-end scenario, no manual triggerable scenario, does not have a triggerable scenario, or similar absence wording. Omit the surface completely unless it actually exists or the user explicitly asks for an explanation of why it is absent.
For every player-facing text surface, the planning spec should define the writing direction and leave finished wording to implementation. This includes event titles, event descriptions, report text, news text, focus text, decision text, option text, achievement text, super-event setup, GUI labels, route flavour, event-detail text, and spreadsheet-facing summaries.
Give the coding agent clear direction for:
Do not provide pasteable localisation. Do not include sample line, possible line, placeholder text, temporary title, or exact draft wording that could be copied into localisation. When a structural label is needed for a file, row, route, branch, asset, or prompt, mark it as a working label, not final localisation.
Text direction should avoid bland map-summary framing and generic communications clichés. Do not make the emotional center of an event a changed map, a line of advance, a staff-table scene, a failed broadcast, or another stock image of crisis administration.
When planning player-facing warning, threat, suspicious-report, and escalation text, do not instruct the coding agent to say that something is a warning, that something is not a warning, that a threat is coming, or that a danger signal has appeared. The text should show the pattern through partial information: fearful witnesses, rumours, anomalies, mysterious powers seen, etc
A report can make the player uneasy without naming the unease. Use mystery, information gaps, fear, rumours, etc. The player should infer that something may be wrong from the content and consequences, not from a label.
When a concept benefits from mystery, fantasy, surrealism, myth, occult signs, prophecy, impossible resolve, strange energy, or unclear public rumours, state that direction clearly without drafting the final prose.
The planning spec may define what an option should feel like, what stance it represents, and how it should vary by route or actor. It must not write final option text.
Do not let the implementation agent default to bland buttons unless plainness is intentional. Describe the intended option style, such as official denial, sarcastic acceptance, bitter understatement, cultural allusion direction, period propaganda tone, frightened understatement, arrogant boast, administrative absurdity, local proverb direction, or grim irony that condemns the speaker.
For each important option or option family, define:
Humour should fit the stakes. Minor chaotic events can use sharper jokes and visible sarcasm. Major disasters, massacres, atrocities, mass death, and real-world suffering should use severity, official euphemism, cynical propaganda, hypocrisy, or self-damning grim irony. Cheap comedy is forbidden there.
Cultural remarks should be treated as research directions unless already sourced. The spec may say that a line should draw from a period slogan, military idiom, folk saying, religious register, old newspaper style, literary echo, propaganda formula, or local public habit. It must not invent the exact remark.
Do not list example option lines. Do not write sample buttons. Do not write placeholder localisation for options. Coding agents may paste those into the game.
Some events need one cutting reaction direction and one plain practical reaction direction. Some need several route-specific reaction directions. The spec should state the intended option tone and purpose, then leave the final wording to the coding agent.
The specification should be as deep as the event idea deserves.
For small events, this may mean a compact but complete spec. For major events, world crises, custom UI systems, event chains, focus trees, custom tags, or events with evolutions, the spec should become very large and multi-part.
Do not aim for a short answer. Aim for a complete design.
For larger events, it is acceptable and expected for the finished specification to span many files and potentially reach tens of thousands or more than 100,000 lines. A huge event with multiple countries, full focus trees, rare variants, and world-order routes can justify 100,000+ lines, and even more, across all parts if the content is meaningful. Do not shorten the spec because it becomes long.
Do not add filler to reach a size target. Add depth by thinking through the event from every useful angle:
Every section should add usable design, player-facing detail, implementation clarity, asset direction, or system connection.
Use research to make the event richer. Do not rely on the first obvious idea.
When the event has historical, cultural, scientific, political, regional, military, religious, or ideological inspiration, research enough to produce specific variants, names, factions, symbols, motives, and consequences.
Research should help answer:
If the topic is niche, current, uncertain, or historically specific, verify with reliable sources. Do not invent source claims. If a source-dependent point is uncertain, mark it as uncertain.
Research should not make the spec dry. Use it to create better event content.
When the event includes decisions, rare variants, country paths, custom actors, or special outcomes, map them out fully.
Do not write vague lines such as:
Instead, define the content. For each meaningful decision or decision group, map:
For rare variants, map:
For branch trees or outcome webs, show the structure. Use headings, named routes, tables, or lists. The coding agent should not have to invent the branch map.
Focus trees must be planned clearly, but the event-planning spec should not try to micromanage every final focus node, every coordinate, or every exact connection. The spec writer should define the tree's routes, branch architecture, major choices, mutual exclusions, story logic, mechanics, rewards, and design standards. The implementation agent should then create the final in-game focus tree layout and exact focus connections cleanly.
A good focus-tree spec should answer:
Do not write only vague branch names. A focus-tree plan must still be detailed enough to prevent generic or boring implementation. The spec should describe the internal logic of each path, the rough order of ideas inside it, its major focus groups, its expected route locks, and its major payoff. It should also name important focuses or focus groups where the story requires them.
Do not require a literal list of every focus unless the user specifically asks for a focus-by-focus blueprint. The default planning style should be path-level and branch-level design. It is acceptable to provide non-final focus-role labels or important anchor focus groups.
Every major focus tree still needs an architecture map. The architecture map should show the intended path structure, not every final focus.
The map should include:
Use a readable structure such as a table, bullet tree, route diagram, or lane map. The implementation agent should understand the intended tree shape and design, but does not need exact focus coordinates from the spec unless the user asks for them.
For each major path, define:
For each important anchor focus or focus group, define:
The coding agent may create more or fewer individual focuses than the spec examples as long as the final tree preserves the path design, story logic, route choices, and gameplay depth.
Focus rewards must be varied. Do not design focus trees where most focuses add a new national spirit, add political power, add stability, add war support, or repeat the same modifier pattern.
A new national spirit or idea should be used only when the focus creates a persistent institution, doctrine, crisis condition, political identity, military structure, economic system, or long-term route effect. If the branch already has an idea representing that institution, prefer modifying, upgrading, replacing, temporarily strengthening, or adding a timed modifier to the existing idea instead of creating another separate idea.
Good focus reward types include:
Every focus path should have a distinct purpose. If two focus groups would grant nearly the same effect, merge them, rewrite one, or make one an upgrade of the other.
Reject focus trees where most focus groups grant new ideas without a clear reason. A tree that uses repeated new ideas as filler has failed even if it has many focuses.
New countries, transformed countries, civil-war splinters, emergency governments, and unstable successor states should not start with a long stack of generic positive ideas. It is usually better for them to start with a small number of deep, readable ideas that define their starting weakness, identity, and strategic problems.
Starting ideas can be negative, mixed, unstable, or conditional. These ideas should represent real problems the country must solve, such as broken administration, improvised command, disputed legitimacy, militia fragmentation, supply confusion, foreign dependence, refugee pressure, factional mistrust, ruined industry, contested railways, disorganized officers, or unclear laws.
Negative starting ideas should not be permanent dead weight unless the story requires that. The spec should map how decisions, missions, focuses, leader choices, foreign aid, victories, reforms, purges, compromises, or crisis outcomes can mitigate, transform, upgrade, or remove them.
Prefer fewer ideas with more depth over many shallow ideas. A good idea can have a lifecycle:
When designing focus trees, do not create a new idea in every focus. If an institution already exists as an idea, prefer changing that existing idea through staged upgrades, replacing it with a route-specific version, adding a temporary modifier, unlocking decisions tied to it, or changing how it interacts with missions and crisis values.
For every important idea, define:
Every major country package should include a starting idea plan and an idea lifecycle table. The table should show which ideas exist at start, which are unlocked later, which are upgraded, which are removed, and which are route-specific.
Example table:
| Idea | Start or unlock | Starting role | Mitigation path | Upgrade path | Failure path | Final forms |
|---|
Reject specs where a country starts with too many unrelated ideas, where every focus creates a separate idea, or where negative ideas cannot be meaningfully addressed through play.
When planning a major focus tree, define more than politics and industry. A large tree needs a distinct expansion, reunification, liberation, settlement, or regional ambition branch. This branch should be separate from the main political tree and separate from the industry tree.
Expansion branches should define real strategic effects, such as claims, cores, war goals, protectorates, guarantees, declarations, leagues, border settlements, ultimatum decisions, or postwar integration choices. Do not reduce expansion to generic bonuses.
Political branches should change politics directly. Define ideology paths, ruling party shifts, party popularity changes, leader changes, advisor unlocks, advisor discounts, laws, councils, juntas, congresses, committees, faction struggles, cosmetic names, and flag changes where they fit. Leader changes imply portrait needs. Real leaders need sourced portraits. Fictional leaders and symbolic councils can use generated portraits through the asset skill.
Fixed-purpose chaos countries can have narrower politics when their identity demands it. A death-state, plague-state, machine-state, or pure destruction actor may have one ideological purpose. Even then, the tree should still provide meaningful internal choices inside that purpose, such as doctrine, expansion method, recruitment, economy, hierarchy, or endgame ambition.
Focus trees and decision systems must be planned together. Focuses should unlock or change decisions and missions. Industry focuses can unlock construction decisions. Military focuses can unlock unit, depot, border, or offensive missions. Diplomacy focuses can unlock recognition, aid, volunteer, and influence decisions. Expansion focuses can unlock declarations, league votes, protectorate demands, border incidents, claims, cores, war goals, and settlement decisions.
For each major focus path, describe which decision or mission families it unlocks and how those decisions expand the mechanic.
Political, industry, and expansion are the minimum branch families, not the full design for important countries. Important countries should usually also define military, diplomacy, internal faction, intelligence or security, special mechanic, and late-game branches when their identity supports them.
Branches should not be isolated columns. Political choices should change which expansion, industry, military, diplomacy, and decision paths are available. Industry should support military or expansion. Expansion should create political consequences. Diplomacy should affect foreign aid, war options, faction choices, and sponsor risk.
Every major branch needs a clear payoff. A political branch can end in a new government, leader, ideology, law system, ruling party, or country identity. An industry branch can end in a rebuilt economy, arsenal, resource system, railway authority, construction mechanic, or production network. An expansion branch can end in a league, empire, federation, protectorate network, reunification, liberation order, regional settlement, or external war plan.
A good focus path should unlock new gameplay, not only stats. The plan should describe decisions, missions, units, advisors, leaders, laws, claims, cores, war goals, buildings, events, mechanics, route access, and AI behavior where they fit. Flat modifiers are supporting rewards, not the main design.
Political routes should update the visible country package where relevant: leader, leader portrait, advisor roster, high command, ruling party, party names, ideology drift or swap, cosmetic name, flag, ideas, and AI strategy. Leader changes imply portrait needs.
Expansion branches should create consequences. Claims, cores, and war goals should interact with diplomacy, factions, resistance, foreign guarantees, local leagues, legitimacy, threat, or postwar settlement decisions.
Industry branches should create map or production changes. Define factories, infrastructure, railways, supply hubs, forts, anti-air, airbases, dockyards, resources, production lines, or construction decisions.
Decision categories should evolve with focus progress. Early focuses may unlock basic decisions. Later focuses should add new targets, stronger actions, cheaper costs, new risks, or new mission families. A decision category should feel different after a route develops.
The fixed-purpose exception is narrow. A country is fixed-purpose only when its concept clearly cannot support normal politics, such as a death-state, machine-state, plague-state, or pure destruction actor. It still needs meaningful internal branches around method, hierarchy, economy, recruitment, expansion, and endgame.
A branch does not count as real unless it changes gameplay. In the spec, each major branch should have several focus groups, a mechanical unlock, a route consequence, and an end-state or payoff.
Major routes need route-specific AI plans. Do not let the implementation use generic focus weights for every route. The spec should say which AI types choose each route and when they avoid it.
Major routes need distinct localisation tone. A socialist route, military route, democratic route, nationalist route, religious route, machine route, death-state route, or foreign client route should not read like the same generic branch with different rewards.
Expansion branches should include postwar handling. War goals alone are not enough. Define claims, cores, puppet options, protectorates, occupation decisions, integration missions, border settlement events, resistance risks, diplomacy reactions, faction consequences, or achievement tracking.
Industry branches should be geographically grounded where possible. Define which states or regions receive factories, resources, ports, railways, supply hubs, forts, anti-air, dockyards, airbases, or infrastructure.
Advisor unlocks should match route identity. Political routes unlock ideological and government figures. Industry routes unlock engineers and economic boards. Military routes unlock commanders and high command. Diplomacy routes unlock envoys and foreign liaisons. High-chaos routes unlock route-specific councils, symbolic leaders, or strange authorities.
Large focus trees should include achievement hooks for difficult route completions, rare branch combinations, expansion outcomes, internal reform, avoiding foreign dependency, league formation, high-chaos survival, or late-game ambitions.
The implementation handoff should require a route coverage table comparing required routes against implemented routes. Missing, renamed, merged, simplified, fallback, or replaced routes must be reported.
A major route should leave visible evidence in the game. The spec should describe what the player actually sees or gains: map changes, decisions, units, advisors, leaders, flags, cosmetic names, faction behavior, focus availability, diplomacy, or visible mechanics. A route that only changes hidden variables or tiny modifiers is not meaningful.
Large focus trees should have early, middle, and late pacing. Early content solves survival and basic identity. Middle content creates route mechanics and real choices. Late content delivers major payoffs, expansion, faction or League outcomes, high-chaos routes, postwar settlement, or world-order ambitions.
Every major route should have a tradeoff. The spec should define what the route risks or sacrifices. Military routes may reduce freedom or legitimacy. Foreign-aid routes may create dependency. Expansion routes may create backlash. Industry routes may consume civilian capacity or weaken short-term defense. High-chaos routes may gain power while damaging stability, diplomacy, or normal politics.
Do not overuse mutual exclusions. Mutually exclusive paths should represent real identity changes, strategic commitments, or incompatible institutions. Support branches such as industry, army, diplomacy, and logistics should usually coexist unless the route logic says otherwise.
Important routes should define failure states. A failed political reform can empower radicals. Failed expansion can trigger backlash or settlement. Failed industry can create dependency. Failed foreign-aid balancing can create a client state. Failed military centralization can create rogue generals or militias.
Focus and decision localisation should describe the visible baseline effect of the route or action. It should not reveal hidden effects, secret outcomes, hidden variables, or future surprises. The player-facing text should explain the public action and visible direction, not the hidden implementation.
Large events should usually include at least one special mechanic. A special mechanic can be a pressure meter, influence system, balance of power, faction cohesion system, legitimacy system, corruption system, outbreak tracker, coalition command system, resource race, regional authority map, or similar play layer.
A special mechanic should define its important values clearly. Examples include legitimacy, authority, influence, cohesion, obedience, corruption, foreign penetration, military readiness, industrial capacity, public panic, faction unity, sponsor pressure, religious authority, revolutionary zeal, or regional control.
Mechanic values must be dynamic. They should move through focuses, decisions, missions, events, wars, state control, foreign influence, AI actions, and prior outcomes. Do not design a mechanic where values only drift passively or change through a few flat scripted effects.
Every important mechanic value should have a consistent colour identity in localisation. If several values contribute to a total, each contributing value should use its own colour consistently across tooltips, scripted localisation, decision text, event text, and UI summaries. If a mechanic has a total value made from components, the tooltip should show a readable breakdown with named and coloured components.
If a mechanic has values such as legitimacy, authority, influence, cohesion, obedience, power, or readiness, then focuses, decisions, and missions should interact with those values directly. A focus tree should not sit beside the mechanic without changing it. A decision system should not sit beside the mechanic without changing it.
Mechanic values should unlock or block content: decisions, focuses, events, missions, leaders, advisors, factions, war goals, reforms, crises, achievements, super-events, or endings. A mechanic should change what the player can do.
When a country has two or more internal power centers, consider a balance-of-power or equivalent system. Focuses and decisions should push the balance, unlock branch content, create risks, and change leaders, laws, advisors, events, or crises.
When an event creates a faction, league, bloc, coalition, compact, or alliance, define its goals and internal rules. The faction should have a reason to exist, membership rules, joining conditions, refusal logic, expulsion or removal logic where relevant, war goals, shared decisions, AI behavior, victory conditions, and failure conditions.
Important event-created factions should usually have a mechanic such as cohesion, shared command, war council support, joint reserves, recognition, member confidence, sponsor pressure, or strategic goals. Focuses and decisions should interact with that faction mechanic.
A faction should not form just because one country exists. Define minimum membership, crisis conditions, ideological compatibility, war pressure, diplomatic preparation, and regional logic.
A special mechanic should define success, failure, partial success, and runaway failure states. These states should unlock events, decisions, focus branches, faction changes, wars, reforms, aftermath, achievements, or super-events.
AI must understand mechanic values. It should know when to lower threat, build legitimacy, increase influence, join a faction, avoid dependence, push balance of power, or trigger escalation.
For every special mechanic, the completion report should list mechanic values, what changes them, what they unlock, UI and localisation coverage, AI behavior, focus hooks, decision hooks, event hooks, and balance checks.
Every special mechanic should define where the player sees it: decision category header, custom scripted GUI, progress meter, scripted localisation tooltip, focus tooltip, national spirit tooltip, or another clear presentation surface. Important mechanic values should not exist only as hidden variables.
When a special mechanic uses a scripted GUI, consider visual presentation beyond static text. Useful designs can include progress bars, meter fill variants, state icons, status frames, warning frames, selected or locked variants, animated frames, or frame-by-frame visual changes that make the mechanic feel alive. The visual layer should make the mechanic easier to understand.
Special mechanics can hide future surprises, but they should not hide basic cause and effect. The player should understand why a visible value rose or fell, which public action changed it, and what kind of response is available.
Faction, league, bloc, or coalition goals should have rewards and failure states. A successful faction goal can unlock shared decisions, war goals, legitimacy, cohesion, member rewards, or postwar settlements. A failed goal can reduce cohesion, trigger exits, invite foreign pressure, start leadership contests, or weaken shared defenses.
New playable country packages must not be generic. Each needs a specific identity, starting problem, political direction, map role, military style, economy, diplomacy, AI behavior, and at least one mechanic or decision family that makes it play differently.
AI strategy must respect route validity. AI should not pick a branch or action that requires a missing state, dead sponsor, non-existent faction, unavailable ideology, disabled evolution, impossible border, invalid target, or absent enemy. Invalid routes should be hidden, bypassed, or weighted to zero.
When a route changes leader, ideology, faction, cosmetic name, flag, advisor roster, or special mechanic identity, define the needed visible assets and whether they are sourced, generated, reused, or blocked.
Shared trees are allowed, but they must have country-specific localisation, route names, decisions, AI weights, leaders, rewards, icons, and scripted localisation where relevant. A shared tree fails if every country using it reads and plays the same.
Important mechanic thresholds, caps, gains, losses, duration bands, AI weights, and scaling values should be centralized in script constants or a clearly documented tuning file. Do not scatter magic numbers across decisions, events, focuses, scripted effects, and scripted triggers.
Avoid one-time reward dumps as the main design. A focus, decision, or mission can give factories, units, equipment, resources, buildings, or influence, but important content should often unlock a repeatable decision, timed mission family, production route, advisor, mechanic, route branch, or long-term gameplay system.
One-time rewards are acceptable when they fit the story and balance. They should not become the default design pattern for a major event, large focus tree, or decision system.
Balance planning should include exploit checks. Look for free unit loops, repeated factory rewards, cheap construction loops, equipment farming, influence farming, puppet abuse, war-goal spam, claim or core spam, advisor discount stacking, bypass abuse, mission success farming, and decisions that can be clicked without meaningful cost or risk.
The spec should tell the implementation agent how to prevent abuse with flags, cooldowns, dynamic costs, escalating costs, one-time completion flags, route locks, target limits, AI limits, cleanup effects, or scripted triggers.
Large decision systems should not show every possible decision at once. The spec should define how decision categories stay readable.
Use phases, caps, priorities, regional pools, route locks, mechanic thresholds, or crisis-state filters so the player sees decisions that matter in the current situation.
Good planning patterns include:
A decision category should feel curated by the current route and campaign state, not like a debug menu.
The implementation agent is responsible for the final exact focus tree shape unless the user asks otherwise.
The implementation agent should:
mutually_exclusive blocksThe spec should give enough creative and structural direction that the agent cannot make a shallow generic tree, while still allowing the agent to build a clean in-game layout.
Focus tree visuals should help the user and implementation agent understand the intended branch structure. The spec may include a high-level branch diagram, lane map, or route sketch for major trees, but it should not try to lock every final focus coordinate unless the user explicitly asks for that.
A useful focus tree visual should show:
The visual should be readable, symmetrical where possible, and free of tangled connector lines. It should not contain random crossing lines or misleading geometry.
If the spec creates a graph or diagram, it should be treated as a design guide unless the user says it must match the final in-game tree exactly. The implementation agent may adjust the final layout to make the actual HOI4 tree cleaner.
Do not spend excessive planning effort forcing exact graph coordinates if the result becomes ugly, brittle, or unhelpful. A clear path architecture is more important than a fake exact graph.
Achievements are mandatory for event specifications unless the user explicitly says not to include them or the event is so small that achievements would be dishonest. Major events, custom countries, deep focus trees, rare variants, world-order routes, or super-events always need achievements.
Achievements should be creative and difficult. Do not design achievements that unlock just because the event fired, because the player clicked the obvious option, or because a country survived a few days. Achievements should reward mastery, unusual campaign states, risky choices, hidden routes, hard containment, difficult victories, or rare evolved outcomes.
A good achievement should usually require several conditions at once, such as:
Do not make achievements conservative. If the event has dark paths, high-chaos tags, strange mechanics, foreign influence systems, coalition politics, or world-order ambitions, design achievements for them. Difficult achievements can require long campaigns, high chaos, multiple wars, internal crises, and careful decision play.
For each achievement, define:
Achievement design must include asset planning. Each achievement needs a 64x64 completed icon direction, and the asset brief must hand those icons to chaos-redux-event-assets. Grey, locked, and not-eligible variants can be produced later if the achievement system requires them.
The achievement list should include a spread of routes. Do not put all achievements on the safest or most obvious path. Cover containment, failure recovery, republic victories, foreign influence, special factions, strange countries, high-chaos routes, and secret or hard branches when those exist.
If an event creates many playable tags, design achievements for the most important ones and for the event-wide systems. A large event can justify dozens of achievements. The achievement handoff should still explain which ones are highest priority if implementation must be staged.
Baseline stages and evolutions are different.
Baseline stages describe the ordinary flow of the event. They are the expected crisis lifecycle, such as first outbreak, containment attempt, spread, coalition formation, deep collapse, settlement, or defeat.
Evolutions are mutation tracks layered on top of the baseline. They make the event more predictable in some ways, more severe, stranger, more patterned, or more replayable. They can add new actors, new rules, new incidents, new tags, old movements, strange variants, stronger breakaways, or rare side branches.
Do not log ordinary stages as evolutions.
Do not use chaos tiers as simple walls that lock ordinary stage progression. Ordinary stages should flow from the event state. Chaos should affect intensity, probability, severity, weirdness, and opening strength.
When an event has evolutions, the spec must say how each evolution enters play. Do not write evolutions only as future modifiers or only as post-fire upgrades unless that is truly the design.
Use two separate entry-path concepts when the event supports both.
Active-event evolution means the event has already fired and an evolution unlocks while one or more actors from that event still exist or the event system is still active. The spec must define what changes immediately for the active actors: focus paths, decision families, national spirits, unit growth, targeting rules, AI strategy, faction behavior, super-event eligibility, and cleanup. The event should not need to fire again for the active actor to receive the evolution content.
Pre-fire evolved opening means the event has not fired yet, but the world state, chaos tier, previous evolution memory, or other allowed event trigger lets the first firing start in a more evolved form. The spec must define the changed opening package: number of actors, target selection, starting ideas, initial units, first decisions, opening events, AI plan.
If both entry paths exist, write both explicitly under one evolution. For example, an event may have an active evolution that unlocks a focus path for an already spawned country, while a later first firing may start with multiple spawned countries.
Each chaos tier can have only one evolution stage. The maximum amount of evolutions is 5.
Good evolutions can include:
Each evolution should define:
Everything that acts like pressure, cooldown, progress, chance, support, duration, cost, tempo, AI willingness, spawn strength, aid amount, stage movement, or recognition should be dynamic by default.
Avoid fixed values as design answers. A fixed number may exist as a tuning anchor, but the spec should define the factors that shape it.
Dynamic factors can include:
Do not say only that a cooldown is 30 days or a pressure increase is 5. Say what makes it shorter, longer, stronger, weaker, safer, or more dangerous.
Dynamic behavior should still be readable. Define cause and effect clearly so the player can learn the pattern through events and decisions.
Political power and command power are useful, but they are usually the least interesting costs. Do not let major decisions, missions, focuses, or crisis responses become a long list of political power and command power purchases.
A good cost should express what the country is actually spending, risking, or sacrificing in the story. A military crackdown may spend command power, but it should also strain units, consume equipment, lower stability, damage war support, pull divisions away from another front, increase resistance, or worsen a crisis pressure. A foreign intervention may spend political power, but it should also require relations work, liaison access, convoys, fuel, equipment shipments, intelligence exposure, or patronage risk. A mobilization decision may use manpower, infantry equipment, support equipment, training time, army XP, supply, local support, or legitimacy.
When mapping costs, use a varied cost palette where it fits the mechanic:
Political power or command power may still be one part of a cost, but they should not be the default answer. If a section uses mostly political power or command power, redesign it unless the story clearly demands bureaucratic or command attention.
Costs should be dynamic. The amount and type of cost should react to country size, chaos tier, stability, war state, equipment stockpiles, supply state, front pressure, foreign access, local legitimacy, previous choices, AI situation, and current event pressure. A weak country should not pay the same cost as a strong country when the story says the burden is different.
Map blocked localisation for nonstandard costs. The player should understand whether they lack infantry equipment, support equipment, divisions in the right state, local support, army XP, fuel, rail control, relations, foreign route access, or another requirement.
For every major decision family, include at least one cost or requirement that is not political power or command power unless the spec explains why that family is purely bureaucratic.
Major event specs must include a real AI section. Do not leave AI behavior as a vague note that the coding agent can decide later.
The AI section should map how every important affected country behaves across the event. This includes the event owner, breakaways, custom tags, transformed existing tags, foreign sponsors, faction leaders, nearby countries, rivals, allies, and countries that can exploit or contain the event.
For each important AI actor or actor group, define:
For focus trees, the spec must define AI path behavior at the branch level and for key individual focuses. If a large tree has mutually exclusive paths, secret routes, or dangerous high-chaos paths, specify which AI personalities or campaign states can choose them. High-chaos AI should be allowed to make strange or extreme choices when that is the point, but ordinary AI should not accidentally choose suicidal or nonsensical branches.
For foreign influence mechanics, the AI section must explain how major powers decide whether to recognize, fund, arm, infiltrate, puppet, betray, or abandon new countries. If volunteers, expeditionary support, proxy wars, or faction invitations exist, AI behavior must be mapped for those too.
A good AI section should make the implementation agent unable to create generic AI weights while claiming to follow the spec.
When an event creates, releases, transforms, or significantly modifies a country, the spec must define that country as a full package. This applies to new custom tags and to existing countries that gain event-specific political identities, focus trees, flags, leaders, cosmetic names, ideology names, starting forces, or mechanics.
For every new country, and every existing country that is meaningfully changed, the spec should provide a country package matrix or equivalent structured section. It must cover:
Do not treat a custom country as complete because it has a tag and one flag. A serious country needs identity, politics, names, flags, leaders, starting forces, force-growth routes, mechanics, decisions, AI, localisation, assets, and route changes. If the country is only temporary and does not need a full package, the spec must explain why.
Political identity should be dynamic when the content supports it. Focus routes, ideology changes, coups, faction victories, foreign puppeting, religious transformations, high-chaos mutations, monarchist restorations, military takeovers, revolutionary councils, or world-order paths should be able to change the country name, flag, ruling party, leader, leader portrait, leader trait, cosmetic tag, national spirits, available decisions, and available recruitment systems when appropriate.
Country names must be direct public country names that remain readable on the map.
Do not build public country names from internal political attachments. Avoid names that use terms such as Military Office, Compact, Bureau, Authority, Mission, Board, or similar administrative labels as the country name. These terms can exist as mechanics, focus groups, advisors, decisions, councils, ministries, internal institutions, route labels, or faction mechanics when they fit. They should not be the public name printed on the map.
Ideology-specific country names are allowed when they fit the route. Sultanate, Kingdom, Empire, Republic, Union, Commune, Federation, and similar public state forms are valid when they match the ideology, historical claim, route, or formable identity.
Prefer names built from the country, people, dynasty, region, or formable identity, then add a simple public state form only when it improves clarity. Style examples include:
AsanteKingdom of AsanteAsante RepublicSultanate of KilwaKongoKongo CommuneDo not overload country names with political office language. The political office, emergency cabinet, military committee, colonial mission, compact council, bureau, authority, or board can be a mechanic or institution inside the country package. The map name should stay short and readable.
Names may depend on ideology, route, leader, formable status, puppet status, or high-chaos transformation. Even then, the public name should remain a country name, not an agency name.
For alternate governments, design internal bodies and party names separately from public country names. A route can have named councils, committees, directorates, juntas, congresses, restoration offices, cult offices, leagues, syndicates, ministries, synods, communes, or military commands. Those institution names should fit the country story, region, history, route, and ideological language without replacing the public country name.
When an event creates, transforms, releases, or empowers countries, check whether formable nations should be part of the design. A formable is a meaningful country identity that appears after a country satisfies territorial, political, event, focus, or hidden-route requirements. Do not treat formables as only a cosmetic rename.
Formable public names must follow the country naming rules above. The formation decision, congress, authority structure, charter, settlement, or council can have its own institutional name, but the formed map name should stay a direct country or formable name.
A formable design should define:
Do not write vague lines such as can form a greater country. Define the concrete formation web. If the player must control this state, this state, and this state, name those states or name the scripted state group and explain what it contains. If the exact state ids are left to implementation, describe the intended geographic set clearly enough that the implementation agent can build a scripted trigger without guessing.
Hidden formables should still be designed fully. The spec can hide player-facing names and spoilers, but the implementation handoff must describe the unlock route, required flags, reveal event, decision visibility, AI behavior, rewards, assets, and disqualifiers.
Formation routes should interact with focus trees and decisions. A focus can reveal or prepare the claim, while a decision performs the formation after the map requirement is met. A decision can form the country directly, while later focuses stabilize it, core it, claim further territory, or resolve internal factions. Avoid giving a formable through a focus alone when the player should prove control over named land first.
When an event creates, releases, transforms, restores, or revives a country that is expected to fight, survive, defend itself, or matter militarily, the spec must define its starting forces. Newly appearing countries should not spawn as empty tags unless they are explicitly non-military administrative placeholders and the spec explains why.
Starting units must be dynamic. Do not define one flat number of divisions for every country. The spec should explain what makes the starting force stronger, weaker, larger, smaller, better equipped, more irregular, more professional, more defensive, more foreign-backed, or stranger.
Useful scaling factors include:
For every meaningful new or transformed country, map:
The spec must also map how newly appearing countries can get more units after spawning. This should include decisions, timed objectives, focus rewards, volunteer systems, depot captures, foreign missions, local mobilization, League or faction training, factory guard mobilization, border guard formation, or special high-chaos recruitment where appropriate.
Do not make reinforcement depend only on political power or command power. Use concrete goals and resources such as holding a capital, guarding a border, controlling a depot, controlling rail lines, spending army XP, consuming equipment, committing manpower, using fuel or trains, securing local support, opening a foreign corridor, finishing a construction quota, proving legitimacy by a deadline, placing divisions in required states, or keeping a volunteer route open.
Unit-creating focuses and decisions must be specific. Avoid repeated generic rewards such as add two infantry divisions across many countries. A unit reward should explain the institution and story behind the unit, such as capital defense committees, local garrison defections, railway guards, factory guard shifts, mountain pass detachments, Black Banner columns, sailor battalions, Basmachi cavalry, ancient host militias, medical volunteers, foreign-trained cadres, or high-chaos special units.
Each unit-creating focus or decision should define:
For focus trees, military growth should be integrated into branches. Some focuses can spawn units directly, but others should unlock decisions, improve templates, recruit commanders, create volunteer corridors, integrate militias, convert irregulars into regulars, expand special units, or change mobilisation rules. A deep tree should offer different ways to build an army depending on politics, foreign influence, economy, terrain, ideology, and chaos state.
Everything visible or meaningful needs an asset plan. A major spec should not only define a few event pictures. It should identify assets for countries, focus trees, decisions, ideas, national spirits, achievements, flags, portraits, faction emblems, super-events, event pictures, UI, unit systems, and route-specific identity changes.
Every focus in a mapped focus tree needs an icon direction. Large trees may use reusable icon packs, but the spec must still state which focuses use which motif or icon category. Do not leave hundreds of focuses with no asset guidance.
Every decision, decision category, idea, national spirit, achievement, faction emblem, UI panel, news image, report image, super-event image, leader or council portrait, and important special-unit identity that appears in the event needs an asset entry or a clear asset-family entry.
Every country package must include flags. Required flag coverage includes normal, medium, and small sizes for each implemented flag state. If the country has ideology-specific names, focus-tree transformations, puppet identities, restored historical forms, radical routes, or high-chaos mutations, the spec must identify whether those states need separate flags.
Historical and real-world flags should not be invented with $imagegen by default. If a country, movement, party, military authority, or restoration path has a real historical flag or a well-attested symbolic design, the asset brief should instruct the asset agent to source that flag or symbol from a reliable source, document it, and convert it into HOI4 flag sizes. Use $imagegen only for fictional, alternate, supernatural, or deliberately invented flag identities when generated art is appropriate.
Historical or real leaders should not be generated. The spec should identify likely real portrait needs and instruct the asset agent to source real images, document source and license status, and crop them to HOI4 portrait size. Fictional leaders, council portraits, cult leaders, alternate invented officers, and symbolic committee portraits can use $imagegen when generated art is appropriate.
When an asset source is historically sensitive, disputed, or politically loaded, the asset brief must require source notes, and a clear distinction between sourced historical use and fictional alternate-history invention.
Do not design important event effects with timid, decorative, or micro values. If an idea, decision, focus, mission, national spirit, crisis response, starting debuff, or route payoff is supposed to matter, its effects must be strong enough for the player to feel and plan around.
Micro modifiers do not count as meaningful design. Avoid values such as plus 2 percent, minus 3 percent, or tiny flat changes as the main reward, penalty, starting debuff, crisis modifier, or route payoff. A small modifier can support a larger effect package only when it belongs to a visible stacking system, a frequent tick, or a clearly explained cumulative mechanic. It cannot be the whole design.
Starting negative debuffs must matter. They should create real pressure on survival, production, mobilisation, logistics, legitimacy, command, diplomacy, state control, AI behavior, or player priorities. A starting problem that the player can ignore has failed. The spec should map how the player feels the debuff, why it exists, which choices mitigate it, and what happens if it is left unresolved.
A good effect package should do at least one meaningful thing:
Effects should fit the event story. A desperate military measure should affect units, equipment, losses, command, supply, stability, or war support. A logistical crisis should interact with trains, fuel, depots, supply, equipment, routes, or tied-down units. A legitimacy crisis should affect stability, war support, recognition, internal factions, local support, or authority. A foreign intervention system should create influence, dependence, access, backlash, or diplomatic consequences.
Normal countries still need balance, but balance must come from costs, timing, risks, limits, tradeoffs, counterplay, AI validity, and route locks. Do not create fake balance by making the numbers too small to matter.
Special chaos countries are different. They do not have to be balanced against ordinary countries. If a special chaos country finishes its path, final route, high-chaos transformation, or full mechanic loop, the payoff must be absurd, dangerous, and visibly overpowered when the concept supports it. A completed chaos country can receive extreme buffs, extreme penalties to enemies, impossible-seeming armies, unnatural production, severe combat bonuses, global pressure, map-changing powers, or other absurd effects if the route earned them.
The spec should still prevent accidental exploits for ordinary countries, repeated free rewards, and unintended cross-route stacking. It should not weaken a completed special chaos country into ordinary balance values. The absurdity must be intentional, visible, and connected to the event identity.
The spec should explain why a value is strong, weak, temporary, risky, conditional, escalating, or deliberately absurd. Reject effect plans whose main outcomes are tiny percentage modifiers, flavour-only ideas, harmless starting debuffs, invisible penalties, or rewards that the player would not notice during normal play.
A strong specification usually explores:
When exploring these areas, do not stop at the first obvious answer. Consider multiple versions and choose or describe the strongest ones.
For important events, think through edge cases, country differences, country package completeness, player incentives, AI incentives, abuse risks, pacing, cooldown factors, narrative tone, repeat play, and evolved behavior.
Consider links to existing Chaos Redux systems when they strengthen the event.
Possible links include:
Leave out connections that feel artificial.
Dangerous systems should not reveal themselves too early.
Player-facing escalation text must not label itself as a warning, a non-warning, a threat, a danger signal, or a world-ending risk. Do not tell the coding agent to write text that announces what the player is supposed to infer. Do not frame an event-detail entry, report, news item, tooltip, decision category, focus description, super-event direction, or spreadsheet-facing summary around that direct label.
Use mysterious information, fear, and uncertainty instead. Early information should feel incomplete because people cannot yet explain what is happening.
Do not build mystery from bureaucratic document motifs, archive-style secrecy, diplomatic evasions, or paperwork drama. Avoid staged contrast formulas that make tension from one side saying or seeing something while an official body denies, delays, softens, avoids, or reacts to it. Avoid timing formulas that make one observation happen before an official admission, public reaction, government response, or wider consequence. Describe the observed fear and uncertainty directly.
The player should understand deeper danger through patterns and consequences over time. It should not explain that the content is a warning or reassure the player that it is not one.
Look for design connections the user may not have considered.
Useful connection types can include:
Add these when they make the event stronger.
If the event benefits from custom UI, design one.
Describe what the player sees, how it changes, and what visual assets are needed.
Map the UI states if the UI represents pressure, route choice, threat, stage, faction cohesion, recognition, contamination, loyalty, or any other living value.
When an event has an important mechanic, decide whether the decision category needs a richer scripted GUI or a separate mechanic window. The spec should define the player-facing interface when the system is important enough to manage visually.
For major events, important decision categories, custom mechanic windows, formable routes, high-chaos route reveals, active crisis meters, special leader transformations, faction boards, patron influence networks, or occult and supernatural systems, run an animation planning pass by default. The pass must either define useful animated sprites or state why static presentation is better for that exact surface. Do not skip the question simply because static assets are easier.
A mechanic UI spec should include:
The spec should not make an interactive window for every small modifier. Use custom UI when it improves readability, choice, atmosphere, or management of a living system.