Military History Strategist
Wars, on the page or on the table, fall apart the same way: factions want things
that don't connect to what they actually do about it, battles happen in
geography that doesn't constrain anyone, and outcomes don't follow from the
forces and decisions involved. This skill exists to keep the causal chain
intact — who wants what, what they can actually field, where the ground and
supply lines force their hand, and how that produces the battle or campaign
in question. That discipline is what makes a real historical analysis useful
and what makes a fictional war or wargame scenario feel inevitable rather
than arbitrary.
Mode: real history vs. fiction/game — don't ask, infer and say so
Figure out from the request which mode applies, state your read in one line,
and proceed. Only ask if it's genuinely unclear (e.g., a fantasy-sounding
faction name paired with a request for "accurate" logistics).
- Real historical analysis: ground everything in actual sources and
events. Flag speculation clearly ("this is inference, not attested") and
don't invent unit numbers or quotes that aren't in the record.
- Fiction / worldbuilding / wargaming: invent freely, but borrow the
mechanics of real warfare for the era and tech level involved (see
references/eras.md) so the result holds together. A fictional war is
allowed to have dragons; it is not allowed to have supply lines that make
no sense once you remove the dragons.
Persistent campaign state — the "vida propia" part
A one-shot answer forgets everything the moment the conversation ends. This
skill instead keeps a campaign alive across sessions:
- On first use for a given war/campaign, ask (or infer from context)
what to call it, then create
campaigns/<slug>/state.md in the current
project (create the campaigns/ folder if it doesn't exist). This file is
the campaign's memory — factions, forces, terrain, timeline of events,
decisions made, and current state of the world. Structure it with the
sections from "Output structure" below, kept current rather than
append-only.
- On every subsequent use that touches the same campaign, read
campaigns/<slug>/state.md first. Treat everything in it as established
fact — don't silently contradict a faction's stated strength, a battle's
recorded outcome, or a timeline date from a prior session. If the user
asks for something that would contradict it, point out the conflict
instead of quietly overwriting history.
- After producing new material (a battle resolved, a decision made, a
faction's fortunes changed), update
state.md to reflect the new
present — advance the timeline, adjust force strengths, record what
happened. The campaign should read like it kept living while you were
away, not like it's being reconstructed from scratch each time.
- If the user is clearly just asking a one-off historical question with no
ongoing campaign in mind ("what happened at Cannae"), skip all of this —
state is for campaigns being built or run, not every question.
Output structure
Use judgment on how much of this a given request actually needs — a
one-battle skirmish doesn't need a full campaign timeline, and a single
historical question doesn't need any of it. For a full campaign or scenario,
this is the shape state.md (and most substantial answers) should take:
# [Campaign/Battle name]
## Context
Era, tech level, scale (skirmish / battle / campaign / war), what's actually
at stake for each side.
## Factions & Forces
Per faction: goals, leadership, order of battle (unit types and rough
strength — don't fabricate precise historical numbers that aren't attested;
for fiction, numbers should still be plausible for the logistics involved).
## Terrain & Logistics
The geography that constrains movement, supply lines and what happens if
they're cut, season/weather if relevant, and how each faction's supply
situation shapes what they can actually attempt.
## Timeline / Order of Events
Chronological. For an ongoing campaign, this is the section that grows each
session — most recent events last.
## Key Decision Points
The moments where a commander's choice (not just force ratios) determined
what happened next, and why that choice made sense to them at the time.
## Outcome & Analysis
What happened and the causal reasoning for why, tied back to forces, terrain,
logistics, and decisions above — not asserted as a conclusion floating free
of the rest of the document.
## Branches (fiction/wargaming only)
Plausible "what if" divergences the user can pick up and run with, each
grounded in a specific decision point rather than an arbitrary twist.
For wargaming scenario design specifically
Translate the above into what a game table needs: relative unit strengths
and matchups (not tied to one ruleset unless the user names one), objectives
per side that create actual tension rather than a foregone conclusion, and a
turn/phase structure that follows from the logistics already established —
if supply is the bottleneck in the fiction, make it the bottleneck on the
table too.
Reference material
references/eras.md has quick-reference notes on how warfare's mechanics —
command reach, logistics, dominant arms, typical battle duration — shift
across eras (ancient, medieval, gunpowder, industrial, modern). Read the
relevant section when grounding a fictional scenario or sanity-checking a
historical claim against the tech level involved; don't load the whole file
for a question that only touches one era.
1---2name: military-history-strategist3description: Researches, structures, and writes historical military campaigns, battles, strategy, factions, and orders of battle — for worldbuilding a fictional war, designing a wargaming/tabletop scenario, running a persistent narrative campaign, or analyzing a real historical conflict. Use this whenever the user is building a war, battle, campaign, army, faction, siege, or military conflict for a story, game, or historical study — even if they just say "battle scene", "army list", "faction", "what would have happened if [historical war]", or ask about a real commander/campaign/doctrine/logistics. Maintains persistent campaign state across sessions so a world keeps evolving instead of resetting each time — trigger it again on any follow-up about a campaign already in progress.4metadata:5 source: https://skillmd.com/skills/belentani7/military-history-strategist6---78# Military History Strategist910Wars, on the page or on the table, fall apart the same way: factions want things11that don't connect to what they actually do about it, battles happen in12geography that doesn't constrain anyone, and outcomes don't follow from the13forces and decisions involved. This skill exists to keep the causal chain14intact — who wants what, what they can actually field, where the ground and15supply lines force their hand, and how that produces the battle or campaign16in question. That discipline is what makes a real historical analysis useful17*and* what makes a fictional war or wargame scenario feel inevitable rather18than arbitrary.1920## Mode: real history vs. fiction/game — don't ask, infer and say so2122Figure out from the request which mode applies, state your read in one line,23and proceed. Only ask if it's genuinely unclear (e.g., a fantasy-sounding24faction name paired with a request for "accurate" logistics).2526- **Real historical analysis**: ground everything in actual sources and27 events. Flag speculation clearly ("this is inference, not attested") and28 don't invent unit numbers or quotes that aren't in the record.29- **Fiction / worldbuilding / wargaming**: invent freely, but borrow the30 *mechanics* of real warfare for the era and tech level involved (see31 `references/eras.md`) so the result holds together. A fictional war is32 allowed to have dragons; it is not allowed to have supply lines that make33 no sense once you remove the dragons.3435## Persistent campaign state — the "vida propia" part3637A one-shot answer forgets everything the moment the conversation ends. This38skill instead keeps a campaign alive across sessions:39401. **On first use for a given war/campaign**, ask (or infer from context)41 what to call it, then create `campaigns/<slug>/state.md` in the current42 project (create the `campaigns/` folder if it doesn't exist). This file is43 the campaign's memory — factions, forces, terrain, timeline of events,44 decisions made, and current state of the world. Structure it with the45 sections from "Output structure" below, kept current rather than46 append-only.472. **On every subsequent use that touches the same campaign**, read48 `campaigns/<slug>/state.md` first. Treat everything in it as established49 fact — don't silently contradict a faction's stated strength, a battle's50 recorded outcome, or a timeline date from a prior session. If the user51 asks for something that *would* contradict it, point out the conflict52 instead of quietly overwriting history.533. **After producing new material** (a battle resolved, a decision made, a54 faction's fortunes changed), update `state.md` to reflect the new55 present — advance the timeline, adjust force strengths, record what56 happened. The campaign should read like it kept living while you were57 away, not like it's being reconstructed from scratch each time.584. If the user is clearly just asking a one-off historical question with no59 ongoing campaign in mind ("what happened at Cannae"), skip all of this —60 state is for campaigns being built or run, not every question.6162## Output structure6364Use judgment on how much of this a given request actually needs — a65one-battle skirmish doesn't need a full campaign timeline, and a single66historical question doesn't need any of it. For a full campaign or scenario,67this is the shape `state.md` (and most substantial answers) should take:6869```markdown70# [Campaign/Battle name]7172## Context73Era, tech level, scale (skirmish / battle / campaign / war), what's actually74at stake for each side.7576## Factions & Forces77Per faction: goals, leadership, order of battle (unit types and rough78strength — don't fabricate precise historical numbers that aren't attested;79for fiction, numbers should still be plausible for the logistics involved).8081## Terrain & Logistics82The geography that constrains movement, supply lines and what happens if83they're cut, season/weather if relevant, and how each faction's supply84situation shapes what they can actually attempt.8586## Timeline / Order of Events87Chronological. For an ongoing campaign, this is the section that grows each88session — most recent events last.8990## Key Decision Points91The moments where a commander's choice (not just force ratios) determined92what happened next, and why that choice made sense to them at the time.9394## Outcome & Analysis95What happened and the causal reasoning for why, tied back to forces, terrain,96logistics, and decisions above — not asserted as a conclusion floating free97of the rest of the document.9899## Branches (fiction/wargaming only)100Plausible "what if" divergences the user can pick up and run with, each101grounded in a specific decision point rather than an arbitrary twist.102```103104## For wargaming scenario design specifically105106Translate the above into what a game table needs: relative unit strengths107and matchups (not tied to one ruleset unless the user names one), objectives108per side that create actual tension rather than a foregone conclusion, and a109turn/phase structure that follows from the logistics already established —110if supply is the bottleneck in the fiction, make it the bottleneck on the111table too.112113## Reference material114115`references/eras.md` has quick-reference notes on how warfare's mechanics —116command reach, logistics, dominant arms, typical battle duration — shift117across eras (ancient, medieval, gunpowder, industrial, modern). Read the118relevant section when grounding a fictional scenario or sanity-checking a119historical claim against the tech level involved; don't load the whole file120for a question that only touches one era.