New-Game Knowledge Initialization

Walkthroughs, wikis, manuals, PDFs and install-folder extras

Start narrow, then grow

Build an explicit onboarding wizard rather than an uncontrolled crawler. It proposes sources, previews scope/size/license hints, and asks the player to approve ingestion categories.

Verified Torment Extras inventory

The live example directory exists. It contains Traveler’s Guide/manual PDFs (including a duplicate), Guide/Torment Strategy Guide.pdf, five novella PDFs, concept-art PDF/CBZ, a world-map image, Extras.rar, and many soundtrack/wallpaper/avatar/ringtone assets.

Initial policy: prioritize manuals, strategy guides, novellas, and text-bearing PDFs. Keep audio and decorative images metadata-only unless multimodal indexing is explicitly enabled.

Initialization stages

  1. Detect identity: normalized game_id, executable, install roots, Steam app ID where available.
  2. Scan local extras: approved install subdirectories only. Example candidate: E:/SteamLibrary/steamapps/common/Torment Tides of Numenera/Extras/.
  3. Inventory: PDFs, HTML, TXT, Markdown, images, manuals and guides; show sizes/hashes before extraction.
  4. Official discovery: official site/support/manual/store/community hub links.
  5. Community source proposals: wikis, GameFAQs, Steam guides, Reddit threads—ranked by authority, completeness and spoiler level.
  6. User selection: full ingest, metadata-only, exclude, or defer.
  7. Extract/index: preserve document structure and page/heading provenance.
  8. Generate guide: a concise corpus manifest, spoiler policy, available references and gaps—not a fabricated walkthrough.

Source adapters

AdapterTechniqueGuardrails
Local PDF/manualPyMuPDF text + page boundaries; OCR fallback only when neededUser-approved roots; hash/version; retain page references
Static wikiGTC map/crawl with domain/path limitsBudget, robots/terms, canonical URLs, incremental refresh
Steam guideGTC scrape selected guide URLsAuthor attribution; avoid account/private content
GameFAQs/RedditSelected URLs or bounded search resultsCommunity trust label; timestamps; do not treat comments as canonical
Official mediaManuals, patch notes, developer posts, subtitles/transcripts when legally accessibleType/edition/version tagging

Spoiler-aware metadata

Classify documents/chunks as mechanics, early-game, quest-specific, full-walkthrough, lore, patch/version, and unknown. Retrieval applies the player’s current spoiler policy before content enters the prompt.

PowerShell-first tooling

Acceptance