agentleFS
ratacatratacat/ratacats-skills/skills/duris-group-combat/SKILL.md

duris-group-combat

Skill9 starsChanged 4 months agoMITView on GitHub

Use when designing or operating Duris group combat with tanks, assists, assumed targets, caster memorization readiness, bash risk, tracking enemies, aggro adds, or command spam risk.

The file

We split this skill into the rules it gives an agent

5 rules. About 0 are common, 4 are distinctive, and 1 is very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈4

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 1

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    REQUIRED BACKGROUND: Use duris-game-knowledge for Duris prompt, posture, memorization, flee, aggro, tracking, and spell mechanics.

    line 19
  2. Distinctive:

    When designing from named logs, use this structure:

    line 32
  3. Distinctive:

    After the binding table, use role names.

    line 52
  4. Distinctive:

    After assist/current target exists, prefer targetless current-target commands.

    line 85
  5. Distinctive:

    Before planned combat, clarify:

    line 89

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Duris Group Combat

Overview

Duris group combat is role-based. The controlled character may be any class; bind live names to roles, then reason with self, tank, leader, enemy, and add.

REQUIRED BACKGROUND: Use duris-game-knowledge for Duris prompt, posture, memorization, flee, aggro, tracking, and spell mechanics.

Priority Order

  1. Survival interrupts: self tanking, tank gone, self disabled, lethal HP, hostile add on self.
  2. Command pacing: prevent duplicate casts, repeated offense, flee spam, and chat spam.
  3. Role action: assist, targetless cast, rescue, order follower, heal, attack.
  4. Communication: concise status the group can act on.

Survival interrupts are not spam. If T: becomes self during a cast, do not wait for spell resolution; send one survival command now and parse the result.

Required Output Shape

When designing from named logs, use this structure:

# Duris Group Combat Behavior

## Role Bindings
| Role | Binding |
|---|---|
| self | <controlled character> |
| tank | <designated or prompt tank> |
| leader | <caller/mover> |
| enemy | prompt E after assist |

## Priority
survival interrupts > command pacing > role action > communication

## States
...

After the binding table, use role names. Do not put character names in headings, state names, or reusable rules.

Hard Stops

Reject any design that does these:

Bad patternCorrect pattern
Says mana, OOM, or out of mana for a Duris mem casterSay not memmed, out of memorized spells, or spell slots spent.
Treats queued mem study as readyReady means memorized and usable now.
Uses cast '<spell>' <enemy>, <target>, or $target after assist/prompt target existsUse a targetless current-target command or configured alias, such as mm, when known.
Defers flee/retreat/rescue because a cast is in progressSurvival interrupts override castLock.
Repeats flee every promptSend one escape command, wait for the result, then reassess.
Invents syntax like flee east or grouptellUse known commands only, such as gsay if grouped, or describe communication intent.
Announces every assist, cast, and spell resultCommunicate readiness and danger, not routine rotation.

Core State Model

StateRule
preflightConfirm target, tank, route risk, and readiness. Casters must report whether key spells are memorized.
assistWhen tank engages or caller says go, use assist <tank> if target acquisition is needed.
engagedPrompt shows stable tank and enemy. Role action may proceed.
castingA spell has started. Block added offense and chatter, but not survival interrupts.
tank-lossSelf is tanking, tank is absent/down/sitting at dangerous time, or tank was forced out. Stop offense.
add-dangerExtra hostile attacks self or destabilizes room. Do not retarget merely because something entered.
escapeSend one survival command, wait for result, then evaluate movement, enemies, and route.
recoverRejoin, heal, mem/med, or report not ready. Do not auto re-engage.

Targeting

The tank owns target selection. The controlled character acquires target by assist or by a prompt enemy already tied to the tank. A wandering arrival is just an arrival until it attacks self, attacks the tank, or the tank/caller clearly picks it up.

After assist/current target exists, prefer targetless current-target commands. Store the enemy name as evidence for communication and logs, not as a spell argument.

Readiness Communication

Before planned combat, clarify:

TopicMeaning
targetWhich enemy the tank will engage first.
tankWho or what should hold aggro: player, pet, follower, charm.
readinessWhether each critical role can act now.
caster memWhich key spells are memorized now, not queued.

During combat, communicate only actionable changes: not ready, out of spells, self tanking, tank down, separated, add on me, fled, need rescue.

Role Notes

Casters: do not cast while sitting, resting, sleeping, or on your ass. Use memorized slots only. If exact targetless syntax is unknown, say configured targetless spell command; do not invent a targeted spell command.

Pet handlers: use pets/followers to establish or restore tanking before personal offense.

Tanks/hitters: establish target clearly. If a weaker group member becomes tank, rescue or retank before maximizing damage.

Fleeability

fleeable includes movement points, posture, known exits, doors, route safety, whether flee goes deeper into danger, tank separation, enemy condition, and tracking evidence. Duris flee is random; directed movement belongs after flee succeeds and combat state allows movement.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.