game-remake-research

name

game-remake-research

description

Source-backed analysis / 拆解 of an existing game into a remake / 复刻 teardown, benchmark / 对标, competitor study / 竞品分析, dossier, spec, formula set, experiment plan, handoff bundle, or vertical-slice research pack / 调研包. Use when the deliverable must separate original-game facts, inferences, and remake decisions. Do not use for generic brainstorming, casual recommendations, implementation-only work, or one-off mechanic explanations.

Game Remake Research

Turn an existing game into source-backed remake documentation. Use this skill for remake teardowns, benchmarks, competitor studies, or vertical-slice research packs, not for generic brainstorming or casual comparisons.

Start

• Lock the target game, edition or era, platform set, region, and time slice before writing conclusions.
• If the target is live-service and the user says latest, current, live, today, or similar, browse first and cite exact dates or build windows.
• Treat repo-local GDDs, spreadsheets, and technical docs as target-state inputs, not evidence about the original game.
• Pick one primary archetype lens. Read references/template-selection.md first, then load only one explicit pair unless a secondary lens is clearly required: mmo -> template-mmo.md + metrics-mmo.md, arpg -> template-arpg.md + metrics-arpg.md, roguelike -> template-roguelike.md + metrics-roguelike.md, card -> template-card.md + metrics-card.md.
• Scaffold the output pack before deep research. For packs that will be audited or handed off, pass a concrete --version-scope and usually add --with-support-files.
• Scaffold templates currently support only --language en. Keep the working research pack in English if you still need audit_remake_pack.py, audit_evidence_links.py, build_pack_status_report.py, or build_handoff_bundle.py. If the final deliverable must ship in another language, translate a derived dossier or handoff artifact after the English pack is complete.

python3 "$GAME_REMAKE_RESEARCH/scripts/scaffold_remake_docs.py" \
  --game "MapleStory" \
  --out ./docs/remake-maplestory \
  --archetype mmo \
  --version-scope "KMS baseline as observed on 2026-03-01" \
  --language en \
  --with-support-files

Example assumes $GAME_REMAKE_RESEARCH points at the skill root. If that variable is unset, run the same script from this skill's local scripts/ directory. Add --with-support-files when the pack should later regenerate experiment summaries or run evidence, status, and handoff tooling against structured support data. Add --single-file only when you also want remake-dossier-template.md as a manual one-file writing template.

Workflow

1. Read references/research-workflow.md and lock scope before collecting sources.
2. Use references/role-matrix.md to cover every analysis lens in one coherent pack.
3. Use references/deliverables.md to keep the pack structure and section outputs aligned.
4. End every major section with open questions and the minimum validation step needed to close them.

Load These References As Needed

• Planning and structure: references/template-selection.md, references/research-workflow.md, references/role-matrix.md, references/deliverables.md
• Archetype lenses: references/template-mmo.md, references/metrics-mmo.md, references/template-arpg.md, references/metrics-arpg.md, references/template-roguelike.md, references/metrics-roguelike.md, references/template-card.md, references/metrics-card.md
• Evidence and capture: references/evidence-rubric.md, references/citation-style.md, references/capture-methods.md
• Experiments and metrics: references/experiment-design.md, references/metric-rollup.md, references/experiment-summary.md, references/evidence-link-audit.md
• Finish and handoff: references/pack-audit.md, references/status-report.md, references/handoff-bundle.md

Evidence Rules

• Prefer sources in this order: official sites, manuals, patch notes, developer talks, direct gameplay footage, reputable wikis or datamines, then secondary commentary.
• Stamp important claims as Confirmed, Inferred, or Open.
• In Confirmed Facts and Inferred Model, cite ledger IDs inline with the claim, for example S12 or S12, S18.
• Attach exact dates, versions, regions, and platforms to balance values, drop rules, UI flows, monetization details, and feature availability.
• If internals are unknowable, infer from repeated observation and explain the inference path instead of pretending certainty.
• Keep original-game facts separate from remake decisions.

Scripts

• Core pack flow: scaffold_remake_docs.py, merge_remake_docs.py
• Experiment flow: rollup_experiment_metrics.py, summarize_experiments.py
• Audit flow: audit_evidence_links.py, audit_remake_pack.py
• Review and handoff: build_pack_status_report.py, build_handoff_bundle.py

If the user wants one final document, merge the pack in the order defined by references/deliverables.md.

python3 "$GAME_REMAKE_RESEARCH/scripts/merge_remake_docs.py" \
  --docs-dir ./docs/remake-maplestory \
  --mode full

Common Failure Modes

• Mixing regions, eras, or platforms without labeling them.
• Loading all four archetype templates when one primary lens would do.
• Writing strong conclusions without an experiment plan for the highest-risk unknowns.
• Confusing inferred formulas with confirmed formulas.
• Writing opinions without backlog, risk, or acceptance criteria.
• Describing art direction without asset categories and reuse strategy.
• Describing feel without timing, state, cancel, or camera rules.
• Producing architecture prose without boundaries, hot paths, or tool implications.

Check Before You Finish

• Confirm the user can see which version or build the research refers to.
• Confirm each role's output is reflected in the final pack rather than left as raw notes.
• Confirm formulas include units, variables, and assumptions.
• Confirm asset and animation sections describe production-ready categories, not just adjectives.
• Confirm architecture and engineering sections define boundaries, sequencing, and risks.
• Confirm unresolved gaps include a validation plan.