复制这段 Prompt，发给你正在使用的 AI 助手。
Follow https://skillsmp.com/skill-install/prompt.md to install Agent Skills from https://github.com/dzhng/civsim.正在加载仓库 README…
已展示 23 / 23 个已收集 Skill。
Review changed code for naming, stale references, unnecessary complexity, and comment quality. Use after completing implementation work, before committing, or when the user asks to review or audit code. Enforces at review time the authoring rules of…
原文语言：英语
How to write and iterate on tests in this repo — one test at a time (tracer bullets), fast cargo first, scale 1v1 before armies, control variables, measure mechanisms not noise. Use when adding sim behavior, fixing a red or brittle test, reviewing a test…
原文语言：英语
Architecture rules for GPU rendering with TypeGPU/WebGPU. Use when adding or changing render passes, shaders, frame orchestration, GPU resources, culling, terrain, shadows, or visual performance work.
原文语言：英语
Compare screenshots against the intended design, distinguishing approved references from historical baselines. Use for iterative implementation matching, before/after visual review, objective image telemetry, or checking a lone capture for flat, empty, or…
原文语言：英语
Use the unprimed sub agent as a second set of eyes before accepting visual work — MANDATORY before declaring any user-reported visual bug fixed or claiming a visual change verified; primed eyes pass defects fresh eyes catch.
原文语言：英语
Implement an existing spec through committed passes. Use for long or multi-pass specs that need maintenance checkpoints to periodically clean code, handoffs, priorities, and plan bloat before drift accumulates.
原文语言：英语
Break large features into independently verifiable, human-reviewable slices under a feature directory in specs/. Use for risky or multi-step feature work that needs upfront questioning, API seams, browser-playable checkpoints, HTML visualizations, screenshot…
原文语言：英语
Use when writing or reviewing geometry, simulation, collision, navigation, culling, procedural generation, transforms, or other performance-sensitive algorithms with the npm `math` package, or when the user invokes /math. Not for routine arithmetic or…
原文语言：英语
When a spec under specs/<feature>/ is done shipping (a write-spec build, or any planned task that produced a spec), archive it to specs/done/ and rewrite it from a build-plan into a durable rationale — the why, the principles, the invariants — pointing back…
原文语言：英语
Take, inspect, and maintain deterministic screenshot baselines for visual changes through the shared snapCheck primitive. Use when verifying UI, campaign, battle, model, animation, or rendering changes; adding or re-blessing snapshots; comparing before/after…
原文语言：英语
Render one committed in-game-pitch contact sheet per model class or prop family — the static model-review gate, pixel-regression gated. Use when adding or changing a model, when a roster/model family grows, or when checking that models read at the…
原文语言：英语
Hunt campaign-map screenshots for visual bugs with fresh-eyes vision subagents — misplaced cities (markers/labels floating on water), sea-name labels sitting on land, missing or dead-end roads, jagged faction-color coastlines, detached or colliding labels.…
原文语言：英语
The visual north star for civsim's battle and campaign rendering: a Bronze-Age Aegean / Total War Saga look. Use when adding or changing terrain, sky, lighting, water, models, labels, icons, faction fills, colors, HUD/UI panels, cards, or bars, or any visual…
原文语言：英语
How to change the SIM PHYSICS (how soldiers move, collide, press, hold a line, rout) — the first-principles workflow and the kind of tests that pin it. Distinct from balancing units. Use when David says a behavior looks wrong ("heavy v heavy isn't clean",…
原文语言：英语
How to change the CAMPAIGN strategic layer in crates/campaign — army movement on the road graph, economy, encounters/battle handoff, the commander AI, diplomacy, orders, and the wasm/frontend campaign loop. Use when adding or changing campaign behavior…
原文语言：英语
Review a model's MOTION as a looping, review-only GIF — pose one asset through its cycle with no sim and eyeball the rhythm. Covers the unit animation cycles (walk/run/attack/hit/die per class) today, and is the home for any future single-asset motion…
原文语言：英语
A vibe is one battle matchup filmed as a time series and watched frame by frame — the realism verdict for sim work. Use when authoring or running a vibe scenario, reading a vibe's frames, shipping or refreshing its GIF, or re-blessing vibe baselines after a…
原文语言：英语
Given a price and a plain-language unit description (melee/ranged, shields, light/heavy, mounted, pole-arm, etc.), fill in the full stat block to fit the game's balance and write the matchup tests that pin it. Use when adding or re-pricing a unit class, or…
原文语言：英语
After a change that moved tests, produce a CHANGE LEDGER for David — every test whose behavior moved, each with {test, previous behavior, new behavior, why}, so he can see what the system now does differently and interrogate any line. Use when wrapping up a…
原文语言：英语
Diagnose hard bugs, regressions, and "feels off" behavior. Build a tight red loop before theorizing, hypothesise, instrument the trajectory, fix behind a regression test. For sim work, the judge is realism not a green test — classify regression-vs-fragility…
原文语言：英语
How to write a verification scene — an addressable, individually-runnable case that boots one world (real map or test map) and asserts one or more visual snapshots and/or behavioral checks. Use when adding or restructuring browser verification for the battle…
原文语言：英语
Audit or rewrite AGENTS.md so it holds only lasting principles. Run it only when the user asks for it by name; never invoke it on your own.
原文语言：英语
Use the local Codex CLI as an independent second agent. Two branches — (1) proactively run `codex review` for a second opinion after completing a substantive change, before presenting it as done or committing; (2) delegate a well-defined implementation task…
原文语言：英语
暂时无法加载仓库 README。你仍可浏览下方的 Skill 目录。