modwright 0.1.0

package/knowledge/baldursgate3/scriptextender.console.yaml

@@ -0,0 +1,257 @@
+game: baldursgate3
+topic: scriptextender.console
+title: Script Extender in-game console usage
+facts:
+  - id: bg3.scriptextender.console.input-line-needs-enter-press
+    claim: The SE console window does not accept typed input immediately — pressing Enter once activates
+      the input line, shown by the prompt changing to "S >>"; if console output has scrolled past
+      the input line, the guide's fix is to "repeat this a few times until you can see the S >>"
+      rather than assuming the console is broken.
+    status: community
+    source: bg3-community/Tutorials/ScriptExtender/GettingStarted.md:54-60, §2.2 Activate the input ("To
+      be able to input commands into the console press enter once."; the next line covers repeating
+      it until the `S >>` prompt is visible)
+    tags:
+      - scriptextender
+      - console
+      - input
+      - gotcha
+  - id: bg3.scriptextender.console.p-and-d-print-functions
+    claim: "Two bare (non-Osi-prefixed) global functions are available at the SE console for printing
+      values: _P (plain print — the guide's worked example is _P(\"Hello World\")) and _D (a richer
+      dump/inspect print, used as _D(Osi.GetHostCharacter()) and
+      _D(Ext.Entity.Get(...).CharacterCreationAppearance))."
+    status: community
+    source: bg3-community/Tutorials/ScriptExtender/GettingStarted.md:67-72, §2.3.1 ("_P("Hello World")")
+      and :525-542, §6.1.1 ("_D(Osi.GetHostCharacter())")
+    tags:
+      - scriptextender
+      - console
+      - print
+      - debug
+      - globals
+    detail: "Checked against the source (2026-09-08): the third citation, §7's
+      `_D(Ext.Entity.Get(GetHostCharacter()).CharacterCreationAppearance)`, has been dropped from
+      `source` — §7 is an unwritten \"Work in Progress\" stub and that string appears only inside
+      its \"// TODO\" list. _P and _D themselves are properly worked examples in §2.3.1 and §6.1.1;
+      the Ext.Entity example is not documented, only listed as something to document."
+  - id: bg3.scriptextender.console.osi-functions-callable-bare-or-prefixed
+    claim: At the SE console, Osiris calls are callable both through the Osi table
+      (Osi.GetHostCharacter(), Osi.AddGold(...)) and as bare global functions with the same name and
+      no Osi. prefix (the guide's TemplateAddTo("d32a68ff-...", GetHostCharacter(), 10) example
+      calls GetHostCharacter() bare inside an argument list to the same effect as
+      Osi.GetHostCharacter()).
+    status: community
+    source: bg3-community/Tutorials/ScriptExtender/GettingStarted.md:76-79, §2.3.2 (bare
+      TemplateAddTo/GetHostCharacter call);
+      bg3-community/Tutorials/Items/Add-and-position-items-in-world.md:32-33
+      ("_P(Osi.GetRegion(GetHostCharacter()))" — GetHostCharacter used bare inside an Osi.-prefixed
+      call)
+    tags:
+      - scriptextender
+      - console
+      - osi
+      - globals
+      - calling-convention
+  - id: bg3.scriptextender.console.reset-command-hotreloads-scripts
+    claim: "Typing \"reset\" at the SE console and pressing Enter \"will reset the console and all
+      loaded scripts, allowing you to see instant changes of your code\" — the guide's workflow
+      after every code edit is: save the file, re-open/re-activate the console, type reset."
+    status: community
+    source: bg3-community/Tutorials/ScriptExtender/GettingStarted.md:420-450, §4.5 Testing the symlink
+      ("This will reset the console and all loaded scripts, allowing you to see instant changes of
+      your code.")
+    tags:
+      - scriptextender
+      - console
+      - reset
+      - hot-reload
+    detail: "The same section carries a precondition the drafted fact omitted: reset only picks up
+      edited scripts if the mod folder is symlinked into (or physically inside) the game's Data
+      folder — otherwise the mod has to be repacked to see changes. It also notes some changes only
+      apply on a session reload (F8)."
+    related:
+      - id: bg3.se.console-reset-degrades-osiris
+        detail: Confirms reset's intended use (hot-reloading Lua) from the community side, matching the
+          existing verified caution that the same command degrades Osiris mid-session and should not
+          be used for anything Osi-driven.
+  - id: bg3.scriptextender.console.bare-expression-prints-nothing
+    claim: 'The Script Extender debug console does not echo the value of a bare expression: typing
+      Ext.Stats.Get("X") returns to the prompt with no output whether or not X exists. Wrap the call
+      in _D(...) (or print) to see the result; _D(Ext.Stats.Get("X")) dumps the whole stats object
+      as JSON-like text (ModId, ModifierList, Using, SpellType, every field), and prints nil for a
+      missing entry.'
+    status: verified
+    verified_on: 4.73.98.727 (Patch 8)
+    source: ModWright field verification
+    tags:
+      - scriptextender
+      - console
+      - lua
+      - dump
+      - stats
+    detail: "The missing-entry case was exercised later the same day
+      (bg3.scriptextender.console.ext-stats-get- missing-entry-prints-null): _D prints null, not
+      nil. The S >> prompt is the server context."
+  - id: bg3.scriptextender.console.ext-mod-getmod-info-directory-names-loaded-copy
+    claim: "Ext.Mod.GetMod(uuid).Info, dumped with _D at the SE console, reports the module the running
+      game bound to that UUID: Directory is the module folder actually loaded
+      (Mods/<Directory>/meta.lsx), alongside Name, Author, Description, ModVersion, PublishVersion,
+      PublishHandle and Hash. When several installed paks or loose modules share one UUID, this is
+      the only observable that says which copy won; modsettings.lsx's Folder attribute is the mod
+      manager's last binding and can disagree."
+    status: verified
+    verified_on: 4.73.98.727 (Patch 8)
+    source: ModWright field verification
+    tags:
+      - scriptextender
+      - console
+      - ext-mod
+      - uuid
+      - duplicate
+      - directory
+    detail: Which copy the game picks among same-UUID installs is not yet characterised (an earlier
+      launch bound a different copy with one more mod.io pak present); this fact is about how to
+      observe the choice, not how it is made. Omitting the closing parenthesis
+      (_D(Ext.Mod.GetMod("...").Info) is a Lua syntax error the console reports as "')' expected
+      near <eof>".
+  - id: bg3.scriptextender.console.getstats-unknown-types-on-runtime-32
+    claim: 'Ext.Stats.GetStats(type) on Script Extender runtime 32 (Patch 8) rejects four of the
+      thirteen type names the API document lists — SpellSet, EquipmentSet, ItemGroup and NameGroup —
+      with a console line "bg3se::lua::stats::FetchStatEntries(): Unknown stats entry type: <type>"
+      and returns nothing for them; the other nine (StatusData, SpellData, PassiveData, Armor,
+      Weapon, Character, Object, TreasureTable, TreasureCategory) answer normally.'
+    status: verified
+    verified_on: 4.73.98.727 (Patch 8)
+    source: ModWright field verification
+    tags:
+      - scriptextender
+      - stats
+      - getstats
+      - console
+      - runtime-32
+    detail: The bridge (bridges/bg3-se, Probes.lua STAT_TYPES) now iterates only the nine known types. A
+      stat of one of the four rejected types is therefore not reachable through bg3.stat.exists on
+      this runtime; whether a later extender build accepts them is unknown.
+  - id: bg3.scriptextender.console.ext-stats-get-missing-entry-prints-null
+    claim: _D(Ext.Stats.Get("<name>")) for an entry that is not loaded prints null, so the same one-line
+      check distinguishes present (a full stats object) from absent (null) without any other setup.
+    status: verified
+    verified_on: 4.73.98.727 (Patch 8)
+    source: ModWright field verification
+    tags:
+      - scriptextender
+      - console
+      - stats
+      - missing
+  - id: bg3.scriptextender.console.db-tables-read-via-get-method
+    claim: Osiris databases are readable from SE Lua/console as members of the Osi table with a
+      colon-called Get method, e.g. Osi.DB_Players:Get(nil) to dump every fact in DB_Players — the
+      guide gives this as the pattern for reading any DB_ database by name (Osi.DB_<Name>:Get(...)).
+    status: verified
+    verified_on: 4.73.98.727 (Patch 8)
+    source: bg3-community/Tutorials/ScriptExtender/GettingStarted.md:635-650, §7 Script Extender Functions
+    tags:
+      - scriptextender
+      - console
+      - database
+      - osi
+      - lua
+    detail: 'Checked against the source (2026-09-08): citation does not support the claim: §7 is an
+      unwritten stub ("Work in Progress", then a "// TODO" list); the `_D(Osi.DB_Players:Get(nil))`
+      string appears only as a bullet in that to-do list of things someone still means to document.
+      The page states no pattern, nothing about the `nil` argument, and nothing about other DB_
+      databases. The syntax may well be right — it just is not documented anywhere read here, so it
+      stayed unverified until run at the console. Run 2026-09-09: the DB_Players form works as
+      written and DB_Origins behaves the same, which closes the "another DB_ name" question. The
+      dump is a list of rows, each row a list of column values; a one-row database prints as its
+      single string.'
+  - id: bg3.scriptextender.console.session-loaded-message-confirms-console-ready
+    claim: After loading or starting a save, the SE console prints a session-loaded confirmation line
+      (plus any installed mods' own debug output) before it is meaningful to issue further commands
+      — the guide treats seeing this output as the signal to proceed, not just the console window
+      existing.
+    status: community
+    source: bg3-community/Tutorials/ScriptExtender/GettingStarted.md:44-50, §2.1 Load a save ("When the
+      save is loaded you should see some output in your console, telling you that the session has
+      loaded and other debug statements depending on what kind of mods you have installed")
+    tags:
+      - scriptextender
+      - console
+      - session-loaded
+      - workflow
+  - id: bg3.scriptextender.console.getposition-getregion-for-capturing-placement-data
+    claim: "Osi.GetPosition(target) and Osi.GetRegion(object) at the console are the documented way to
+      capture a level-placement data point without a mod dependency: \"Using Script Extender to get
+      positional information does not create a dependency on it for your mod, you're simply using it
+      to capture data points.\" GetPosition returns X, Y, Z space-separated (no commas) when
+      printed, but the values must be re-entered comma-separated into other Osi calls."
+    status: community
+    source: bg3-community/Tutorials/Items/Add-and-position-items-in-world.md:21 ("Using Script Extender
+      to get positional information does not create a dependency on it for your mod, you're simply
+      using it to capture data points.") and :29-39, §Position (the GetRegion/GetPosition commands
+      and the no-commas-when-printed, commas-when-re-entered note)
+    tags:
+      - scriptextender
+      - console
+      - getposition
+      - getregion
+      - workflow
+  - id: bg3.scriptextender.console.createat-spawns-a-copy-not-the-original
+    claim: Osi.CreateAt(templateId, x, y, z, temporary, playSpawn, spawnFinishEvent) called at the
+      console "doesn't add the original to the map, we're creating a copy based on the original" —
+      templateId is the target RootTemplate's MapKey, temporary=0 makes the spawned copy permanent,
+      and the call returns a GUIDSTRING for the new object that should be captured into a console
+      variable for later reference (e.g. newbox = Osi.CreateAt(...)).
+    status: community
+    source: bg3-community/Tutorials/Items/Add-and-position-items-in-world.md:43-57, §Spawn Item ("It
+      doesn't add the original to the map, we're creating a copy based on the original."; worked
+      example assigning the return to `newbox`)
+    tags:
+      - scriptextender
+      - console
+      - createat
+      - roottemplate
+      - workflow
+  - id: bg3.scriptextender.console.itemrotateytoangle-is-absolute-not-relative
+    claim: Osi.ItemRotateYToAngle(item, angleDeg, speedDegPerSec) rotates an item TO an absolute Y-axis
+      angle, not BY that angle — issuing the same command with 90 degrees twice in a row leaves the
+      item at 90 degrees both times, it does not accumulate to 180. Osi.ItemRotateY (without
+      "ToAngle") is a separate, distinctly named generated call with the same three parameters.
+    status: community
+    source: bg3-community/Tutorials/Items/Add-and-position-items-in-world.md:70-84, §Rotation ("We are
+      not rotating **BY** that angle."; the paragraph then works through the repeated-90-degrees
+      case); LaughingLeader/BG3ModdingTools@ab343b6 generated/Osi.lua:3107 (ItemRotateY), :3112
+      (ItemRotateYToAngle)
+    tags:
+      - scriptextender
+      - console
+      - itemrotateytoangle
+      - rotation
+      - gotcha
+    detail: 'Checked against the source (2026-09-08): an earlier wording called ItemRotateY "the call
+      for a relative rotation". The tutorial never mentions ItemRotateY and the generated stub gives
+      only its name and parameters, so its semantics are an assumption from the name; only its
+      separate existence is cited here.'
+  - id: bg3.scriptextender.console.lua-loader-takes-code-then-an-env-table
+    claim: "A mod's server Lua state under Script Extender runtime 32 exposes a Lua loader (reached as
+      `loadstring or load`) that compiles a source string, and its optional second argument is an
+      environment table, not a chunk name: calling it as loader(code, \"=name\", \"t\") raises \"bad
+      argument #2 to 'loader' (table expected, got string)\", while loader(code) compiles and the
+      chunk runs against the mod's globals (Ext, Osi and the rest resolve)."
+    status: verified
+    verified_on: 4.73.98.727 (Patch 8)
+    source: ModWright field verification
+    tags:
+      - scriptextender
+      - lua
+      - load
+      - loadstring
+      - sandbox
+      - bridge
+      - eval
+    detail: Which of `load` and `loadstring` the state actually binds was not separated (the handler
+      takes the first defined); the second-argument shape matches the classic sandboxed-loadstring
+      convention rather than Lua 5.4's load(chunk, name, mode, env). Osi.TemplateAddTo returns
+      nothing (returnCount 0), so the inventory is the only read-back;
+      Osi.GetItemByTemplateInInventory returns one guid and cannot count.

package/knowledge/baldursgate3/scriptextender.yaml

@@ -0,0 +1,252 @@
+game: baldursgate3
+topic: scriptextender
+title: Script Extender configuration, APIs, and Osiris integration
+facts:
+  - id: bg3.se.mutually-exclusive-with-toolkit
+    claim: SE and the official Toolkit are mutually exclusive for AUTHORING — pick the hand-authored
+      pipeline for SE mods. (Publishing an SE mod through the Toolkit is a separate question; see
+      bg3.publishing.se-mods-can-ship-on-modio.)
+    status: verified
+    verified_on: 4.73.98.727 (Patch 8)
+    source: ModWright field verification
+    tags:
+      - scriptextender
+      - toolkit
+      - authoring
+    detail: This is why the whole BG3 pipeline here is hand-authored + LSLib/Divine + Nexus, with the
+      Toolkit used only to publish to mod.io. It is also why the Toolkit's "Convert UI Assets" step
+      (which would auto-write GUI/metadata.lsf) is unavailable.
+  - id: bg3.se.config-required
+    claim: 'SE config MUST be Mods/<Mod>/ScriptExtender/Config.json — capital C (pak-internal paths are
+      case-sensitive; lowercase config.json silently fails in a packed mod) — with the
+      RequiredVersion key, NOT the DOS2-era RequiredExtensionVersion. Shape: { "RequiredVersion":
+      21, "ModTable": "<Mod>", "FeatureFlags": ["Lua"] }.'
+    status: verified
+    verified_on: 4.73.98.727 (Patch 8)
+    source: ModWright field verification
+    tags:
+      - scriptextender
+      - config
+      - bootstrap
+    detail: "Both failure modes were found on an earlier mod: a lowercase config.json (tolerated for
+      loose files, fatal in a pak) and the DOS2-era RequiredExtensionVersion key — both silent. SE
+      enumeration works loose and packed, but only for mods in the active load order.
+      RequiredVersion is a floor: set it to the lowest SE major that has the APIs you use."
+  - id: bg3.se.bootstrap-files
+    claim: "Server bootstrap: Mods/<Mod>/ScriptExtender/Lua/BootstrapServer.lua (runs when a session
+      loads). Client bootstrap: Mods/<Mod>/ScriptExtender/Lua/BootstrapClient.lua (runs at the main
+      menu)."
+    status: verified
+    verified_on: 4.73.98.727 (Patch 8)
+    source: ModWright field verification
+    tags:
+      - scriptextender
+      - bootstrap
+      - lua
+    detail: The file names and paths are exact. Register listeners on SessionLoaded.
+  - id: bg3.se.osiris-listener-arity
+    claim: 'Listen to Osiris events with Ext.Osiris.RegisterListener("StatusApplied", 4, "after", fn).
+      Get the correct ARITY — wrong arity = dead listener. Signatures: StatusApplied(object, status,
+      causee, storyActionID) [4]; StatusRemoved(object, status, causee, applyStoryActionID) [4];
+      CastedSpell(caster, spell, spellType, spellElement, storyActionID) [5]; KilledBy(defender,
+      attackOwner, attacker, storyActionID) [4]; Died(character) [1];
+      TemplateKilledBy(characterTemplate, defender, attackOwner, attacker, storyActionID) [5].'
+    status: verified
+    verified_on: 4.73.98.727 (Patch 8)
+    source: ModWright field verification
+    tags:
+      - scriptextender
+      - osiris
+      - listeners
+      - arity
+    detail: Canonical signatures come from LaughingLeader BG3ModdingTools/generated/Osi{,.Events}.lua;
+      the 2026-09-08 intake drafted a second fact restating these same six signatures from that file
+      (bg3.osiris.signatures.listener-event-arity-confirmed-against-generated-file), which was
+      deleted as a duplicate with its line citations merged into this fact's source instead.
+      Correlate related events on the shared storyActionID — but its lifetime is UNDOCUMENTED and
+      IDs are reused within a session, so any correlation table needs a TTL/cleanup sweep; never
+      treat a storyActionID as globally unique. A weapon kill carries its own storyActionID distinct
+      from the spell that applied the victim's status.
+  - id: bg3.se.v32-api-date
+    claim: The SE v32 API reference in these notes was verified 2026-08-29 against
+      Ext.Utils.Version()==32; v32 released 2026-06-21. Canonical Osiris signatures from
+      LaughingLeader BG3ModdingTools/generated/Osi{,.Events}.lua; Lua events from Norbyte/bg3se
+      BG3Extender/IdeHelpers/ExtIdeHelpers.lua.
+    status: verified
+    verified_on: 4.73.98.727 (Patch 8)
+    source: ModWright field verification
+    tags:
+      - scriptextender
+      - se-v32
+      - api
+      - reference
+    detail: "The damage pipeline names churn between versions: current ones are
+      Ext.Events.EsvLuaBeforeDealDamage (fires BEFORE mitigation — modify here) and
+      Ext.Events.EsvLuaDealDamage. StatsDealDamageFunctor has IgnoreImmunities but no
+      IgnoreResistances, so resistance bypass must be done in these hooks. Old DOS2-era Game.Math /
+      bare DealDamage names are gone."
+  - id: bg3.se.use-spell-forced-cast
+    claim: Osi.UseSpell(caster, spellID, target[, target2[, withoutMove:int]]) — overloads 3/4/5 args,
+      NO level parameter; Osi.UseSpellAtPosition(caster, spellID, x, y, z[, withoutMove]) for point
+      casts. Forced casts resolve at spell power level 0, so encode any level in the spell ID (cast
+      the pre-leveled child), not the call.
+    status: verified
+    verified_on: 4.73.98.727 (Patch 8)
+    source: ModWright field verification
+    tags:
+      - scriptextender
+      - osiris
+      - usespell
+      - forced-cast
+    detail: SpellPowerLevel-scaled effects therefore see 0 on forced casts. Whether forced casts consume
+      resources was still untested at the time of that finding.
+  - id: bg3.se.console-reset-degrades-osiris
+    claim: 'SE console `reset` hot-reloads BootstrapServer.lua/BootstrapClient.lua without restarting
+      the game, BUT reset mid-session degrades Osiris ("Not syncing adapters - DB not available"):
+      Osi queries return junk and Osi calls silently no-op afterwards. Use reset only to reload Lua
+      code; anything Osi-driven needs a save reload (F8) instead.'
+    status: verified
+    verified_on: 4.73.98.727 (Patch 8)
+    source: ModWright field verification
+    tags:
+      - scriptextender
+      - console
+      - reset
+      - osiris
+    detail: "Degradation detector: Osi.GetMaxHitpoints(host) vs Ext.Entity.Get(host).Health.MaxHp
+      disagreeing. Quickload (F8) starts a fresh session — SessionLoaded re-fires, Osi state is
+      clean, and session-granted passives/statuses are wiped back to the save's state."
+  - id: bg3.se.lua-best-practices
+    claim: Wrap everything in pcall. Remove dev diagnostics (UI-viewmodel probes, StaticData dumps)
+      before release. Stats entries are live-editable from the SE console —
+      Ext.Stats.Get("<entry>"), then SetRawAttribute + Sync; functor/complex fields REQUIRE
+      SetRawAttribute (plain assignment throws "Inappropriate type" in SE v32).
+    status: verified
+    verified_on: 4.73.98.727 (Patch 8)
+    source: ModWright field verification
+    tags:
+      - scriptextender
+      - lua
+      - error-handling
+      - diagnostics
+    detail: Simple string fields accept plain assignment. Stats .txt FILE edits still need a game
+      restart (a whole-file runtime reload API is unconfirmed), and Progressions/ClassDescriptions
+      .lsx are read at session load only.
+  - id: bg3.se.corpse-removal
+    claim: 'Corpse/entity removal: Osi.RequestDelete(obj), guarded with Osi.IsDead(obj)==1 and
+      Osi.IsPlayer(obj)~=1. DeathType "Disintegrate" only drives the dissolve VFX; the corpse entity
+      persists — removing it needs SE.'
+    status: verified
+    verified_on: 4.73.98.727 (Patch 8)
+    source: ModWright field verification
+    tags:
+      - scriptextender
+      - osiris
+      - corpse
+      - deletion
+    detail: Stats-side disintegration is cosmetic only, which is why the visual alone never removes the
+      body.
+  - id: bg3.se.osi-queries-broken-actions-work
+    claim: "In this SE build, Osiris ACTIONS work but Osi QUERIES are unreliable: Osi.SetFlag works
+      while Osi.GetFlag returned 0 for a provably-set flag, and every Osi query is treated as
+      suspect."
+    status: verified
+    verified_on: 4.73.98.727 (Patch 8)
+    source: ModWright field verification
+    tags:
+      - scriptextender
+      - osiris
+      - queries
+      - flags
+    detail: "Design around it: drive state with actions and read state from Ext.Entity / stats rather
+      than Osi queries. Note that a degraded session (console `reset`) produces the same symptom for
+      a different reason — see bg3.se.console-reset-degrades-osiris."
+  - id: bg3.se.osiris-restricted-at-sessionloaded
+    claim: 'Osiris functions are NOT callable during SessionLoaded (and other early load phases):
+      calling any Osi.* there throws "Attempted to call Osiris function in restricted context".
+      Registering listeners (Ext.Osiris.RegisterListener) at SessionLoaded is fine; calling an
+      Osiris function synchronously at that point is not.'
+    status: verified
+    verified_on: 4.73.98.727 (Patch 8)
+    source: ModWright field verification
+    tags:
+      - scriptextender
+      - osiris
+      - sessionloaded
+      - restricted-context
+      - lua
+    detail: Move any actual Osi.* call (GetHostCharacter, GetFlag, TemplateAddTo, ...) out of
+      SessionLoaded and into a gameplay/Osiris event handler where the story is live (e.g. a
+      level-load or region listener). An Ext.Osiris.IsCallable() guard, if present in the installed
+      SE build, can gate calls from ambiguous contexts; deferring to an Osiris event is the reliable
+      fix.
+  - id: bg3.scriptextender.osi-inventory-calls-complete-asynchronously
+    claim: 'Osi.ToInventory, Osi.TemplateAddTo, Osi.TemplateRemoveFrom and Osi.RequestDelete return
+      before the change is applied: a GetItemByTemplateInInventory / GetInventoryOwner / Exists read
+      in the same chunk, or 2-4 s later in a following probe, still shows the old state; 8-20 s
+      later it shows the new one. A read-back must settle for at least ~8 s, and issuing a
+      "fallback" TemplateAddTo because the first read looked unchanged creates a duplicate.'
+    status: verified
+    verified_on: 4.73.98.727 (Patch 8)
+    source: ModWright field verification
+    tags:
+      - scriptextender
+      - osiris
+      - inventory
+      - async
+      - bridge
+  - id: bg3.scriptextender.story-items-are-matched-by-instance-not-template
+    claim: "Osiris goals bind to a specific global item INSTANCE (UseFinished(_Player,
+      S_GLO_PixieBell_8c99b6c7-..., 1) in GLO_Pixie.txt), so a second item spawned from the same
+      template (TemplateAddTo of UNI_SCL_PixieBell's MapKey) is inert: its use action runs but no
+      goal reacts. To hand a player such an item, move the real instance with
+      Osi.ToInventory(\"<Name>_<MapKey>\", host, 1, 1, 1); the story parks such items at a hidden
+      \"asylum\" trigger per level (the pixie and her bell stood at 20/0/-1417 in SCL_Main_A) so the
+      instance exists to be moved."
+    status: verified
+    verified_on: 4.73.98.727 (Patch 8)
+    source: Gustav.pak GLO_Pixie.txt lines 37-42; bridge requests req-mu8oxy6g, req-mu8p0lk3,
+      req-mu8p11jx, req-mu8p3i4c
+    tags:
+      - osiris
+      - story
+      - global-items
+      - templateaddto
+      - toinventory
+  - id: bg3.scriptextender.level-placement-via-levelgameplaystarted-and-container-instance-id
+    claim: "A mod can place its item in a specific vanilla container natively-shaped for SE: register
+      Ext.Osiris.RegisterListener(\"LevelGameplayStarted\", 2, \"after\", fn) at SessionLoaded and,
+      when the level name matches, Osi.TemplateAddTo(<item template>, \"<ContainerName>_<MapKey>\",
+      1, 0) after a GetItemByTemplateInInventory dedup. The container id is the level
+      Items/_merged.lsf node's Name and MapKey joined by \"_\"
+      (CONT_GEN_Rock_Pile_Grave_Open_A_000_472199db-9e39-4f40-9a0d-9c439b1953bf). It fires on a save
+      loaded into that level, not only on first entry."
+    status: verified
+    verified_on: 4.73.98.727 (Patch 8)
+    source: ModWright field verification
+    tags:
+      - scriptextender
+      - osiris
+      - placement
+      - levelgameplaystarted
+      - containers
+  - id: bg3.scriptextender.in-game-container-survey-through-the-bridge
+    claim: 'The reliable way to find "the container the player is standing at" is a bridge bg3.lua.eval
+      on the server: Osi.GetPosition(Osi.GetHostCharacter()), then iterate
+      Ext.Entity.GetAllEntitiesWithComponent("ServerItem"), keep entities whose
+      Transform.Transform.Translate is within a few metres, and print Uuid.EntityUuid,
+      ServerItem.Template.Name, Osi.IsContainer(uuid),
+      Ext.Loca.GetTranslatedString(DisplayName.NameKey.Handle.Handle) and
+      Osi.GetInventoryOwner(uuid) (which shows what already sits inside which container). Offline
+      level surveys anchored on a name-regex centroid picked the wrong object twice (a cellar chest
+      950 m from the graveyard; a chest 45 m from the display case).'
+    status: verified
+    verified_on: 4.73.98.727 (Patch 8)
+    source: ModWright field verification
+    tags:
+      - scriptextender
+      - bridge
+      - survey
+      - entities
+      - containers
+      - levels