modwright 0.1.0

package/knowledge/skyrimse/plugins.format.yaml

@@ -0,0 +1,224 @@
+game: skyrimse
+topic: plugins.format
+title: Plugin file layout — the record header, the TES4 file header and its subrecords
+facts:
+  - id: sse.plugins.format.record-header-is-24-bytes
+    claim: "Every record in a Skyrim SE plugin begins with a 24-byte header: 4 bytes of signature,
+      uint32 data size (excluding the header), uint32 flags, uint32 FormID, 4 bytes of
+      timestamp/version-control info, uint16 form version, uint16 unknown. The record's data follows
+      immediately at offset 24."
+    status: community
+    source: en.uesp.net "Skyrim Mod:Mod File Format" section "Records"; corroborated by TES5Edit
+      wbImplementation.pas (TwbMainRecordStruct) and Ortham/esplugin src/record.rs
+      (MAX_RECORD_HEADER_LENGTH = 24, test header_length_should_be_24)
+    tags:
+      - plugin
+      - binary
+      - record-header
+      - parser
+      - byte-layout
+    detail: UESP splits offsets 16-19 into two uint16 fields (Timestamp, Version Control Info) where
+      xEdit reads one Cardinal (mrsVCS1); both readings parse the same bytes. All multi-byte fields
+      are little-endian.
+  - id: sse.plugins.format.form-version-44-for-sse
+    claim: "The uint16 at record-header offset 20 is the record's internal form version: 43 for Skyrim
+      LE and 44 for Skyrim Special Edition."
+    status: community
+    source: en.uesp.net "Skyrim Mod:Mod File Format", section "Records", the uint16 internal-version row
+    tags:
+      - plugin
+      - record-header
+      - version
+      - sse-vs-le
+    detail: A quick way to tell an LE-authored plugin from an SE one without parsing further. Fallout 4
+      uses 131 on the same axis.
+  - id: sse.plugins.format.tes4-flag-bits
+    claim: On the TES4 file-header record the meaningful flag bits are 0 (0x00000001) ESM master, 7
+      (0x00000080) Localized, 9 (0x00000200) Light/ESL, and 12 (0x00001000) Ignored. xEdit declares
+      only the first three on TES4 and gates the ESL bit on the game being Skyrim SE.
+    status: community
+    source: TES5Edit wbDefinitionsTES5.pas, wbRecord(TES4, 'Main File Header', wbFlags(...)) — bit 9
+      written as IsSSE('ESL',''); matortheeternal/esp.json data/SSE/TES4.json "flags"; en.uesp.net
+      "Skyrim Mod:Mod File Format/TES4" lead paragraph
+    tags:
+      - plugin
+      - flags
+      - esm
+      - esl
+      - localized
+      - parser
+    detail: "The ESM flag's effect, verbatim from the TES4 page: \"if set, the file is treated as a
+      master, regardless of what the file extension says\". The bit values are not portable across
+      games — esplugin's plugin.rs shows Starfield using 0x800 for light and 0x200 for Update."
+  - id: sse.plugins.format.localized-flag-changes-string-decoding
+    claim: The Localized flag (0x00000080) makes the game load the .STRINGS, .DLSTRINGS and .ILSTRINGS
+      files that accompany the plugin. With the flag clear, lstrings in the plugin are treated as
+      plain zstrings instead.
+    status: community
+    source: en.uesp.net "Skyrim Mod:Mod File Format", section "Records", the 0x00000080 flag row
+    tags:
+      - plugin
+      - flags
+      - localization
+      - strings
+    detail: A plugin shipped with the flag set but without its strings files loses every localized name,
+      which is why the flag is worth reporting from a header parse.
+  - id: sse.plugins.format.tes4-subrecord-order
+    claim: The TES4 record's subrecords, in the order xEdit declares them, are HEDR, OFST, DELE, CNAM,
+      SNAM, then repeating MAST+DATA pairs, then ONAM, SCRN, INTV and INCC.
+    status: community
+    source: TES5Edit wbDefinitionsTES5.pas, the wbRecord(TES4, 'Main File Header', ...) member list;
+      reproduced exactly by matortheeternal/esp.json data/SSE/TES4.json
+    tags:
+      - plugin
+      - tes4
+      - subrecord
+      - parser
+      - byte-layout
+    detail: OFST, DELE and SCRN appear in xEdit and esp.json but not on the UESP TES4 page. A parser
+      should skip unknown subrecords by their declared size rather than assume this exact set.
+  - id: sse.plugins.format.subrecord-header-is-6-bytes
+    claim: "A subrecord (field) header is 6 bytes — 4-byte signature plus uint16 data size — with one
+      exception: if the preceding field has the type XXXX, this field's size reads 0 and the real
+      size is the uint32 stored in that XXXX field."
+    status: community
+    source: en.uesp.net "Skyrim Mod:Mod File Format", section "Fields"
+    tags:
+      - plugin
+      - subrecord
+      - parser
+      - byte-layout
+      - gotcha
+    detail: The XXXX escape exists because a uint16 cannot express a field larger than 65535 bytes. It
+      does not occur inside TES4, so a header-only parser can ignore it, but any parser that walks
+      past the file header must handle it.
+  - id: sse.plugins.format.hedr-is-twelve-bytes
+    claim: "HEDR is a required 12-byte struct: float32 version, uint32 number of records and groups (not
+      counting the TES4 record itself), uint32 next available object ID."
+    status: community
+    source: en.uesp.net "Skyrim Mod:Mod File Format/TES4", HEDR row; TES5Edit wbDefinitionsTES5.pas
+      (wbFloat('Version'), wbInteger('Number of Records', itU32), wbInteger('Next Object ID',
+      itU32), marked required)
+    tags:
+      - plugin
+      - tes4
+      - hedr
+      - parser
+      - byte-layout
+  - id: sse.plugins.format.mast-data-pairs-are-the-master-list
+    claim: Each master is a MAST subrecord (a zstring filename) immediately followed by an 8-byte DATA
+      subrecord. The masters are listed in the load order they had when the plugin was saved, and
+      the mod index of every FormID in the plugin refers to that order.
+    status: community
+    source: en.uesp.net "Skyrim Mod:Mod File Format/TES4", MAST row; TES5Edit wbDefinitionsTES5.pas
+      wbRArray('Master Files', wbRStruct('Master File', [MAST, DATA]))
+    tags:
+      - plugin
+      - tes4
+      - mast
+      - masters
+      - formid
+      - load-order
+    detail: DATA is "Always 0. May be vestigial. In Tes3, the file size of the master was recorded
+      here." xEdit models it as an 8-byte array marked cpIgnore but required. Because FormID mod
+      indices are positional, reordering a plugin's masters without remapping its FormIDs silently
+      repoints every reference.
+  - id: sse.plugins.format.author-and-description-are-512-bytes
+    claim: CNAM holds the author and SNAM the description, each a Windows-1252 zero-terminated string
+      with a maximum size of 512 bytes including the terminator.
+    status: community
+    source: en.uesp.net "Skyrim Mod:Mod File Format/TES4", CNAM and SNAM rows; Wrye Bash
+      Mopy/bash/game/__init__.py (max_author_length = 511, max_desc_length = 511)
+    tags:
+      - plugin
+      - tes4
+      - cnam
+      - snam
+      - author
+      - description
+      - limits
+    detail: "The two sources disagree on whether CNAM is required — xEdit marks it required, UESP marks
+      it optional. SNAM is worth reading beyond curiosity: Wrye Bash reads Bash Tags out of the
+      plugin description."
+  - id: sse.plugins.format.onam-only-in-esm-overriding-cell-children
+    claim: ONAM is an array of FormIDs (count = payload size / 4) that appears only in ESM-flagged files
+      which override their masters' cell children.
+    status: community
+    source: en.uesp.net "Skyrim Mod:Mod File Format/TES4", ONAM row ("Observed in Update.esm as of Patch
+      1.5.24")
+    tags:
+      - plugin
+      - tes4
+      - onam
+      - esm
+      - cell
+    detail: UESP lists the overridable types as ACHR, LAND, NAVM, PGRE, PHZD and REFR; xEdit's allowed
+      list is wider, adding PMIS, PARW, PBAR, PBEA, PCON and PFLA.
+  - id: sse.plugins.format.zstring-is-windows-1252
+    claim: A zstring in a plugin file is a zero-terminated string whose size counts the text plus one
+      terminator byte, encoded Windows-1252.
+    status: community
+    source: en.uesp.net "Skyrim Mod:File_Format_Conventions", section "Data Types"
+    tags:
+      - plugin
+      - encoding
+      - zstring
+      - windows-1252
+      - parser
+    detail: The same encoding governs Plugins.txt (see sse.loadorder.files.plugins-txt-is-windows-1252),
+      so a tool that gets one right and the other wrong will corrupt exactly the mods whose names
+      motivated the rule.
+  - id: sse.plugins.format.compressed-record-flag
+    claim: "A record whose flags include 0x00040000 stores its data compressed: a uint32 decompressed
+      size followed by dataSize minus 4 bytes of zlib data."
+    status: community
+    source: en.uesp.net "Skyrim Mod:Mod File Format", section "Records"; Ortham/esplugin src/record.rs
+      tests (self.flags & 0x0004_0000) != 0
+    tags:
+      - plugin
+      - compression
+      - zlib
+      - parser
+    detail: TES4 itself is never compressed, so a header-only parser is unaffected; any parser that
+      reads further must handle it or it will read garbage.
+  - id: sse.plugins.format.grup-size-includes-its-own-header
+    claim: A GRUP header is also 24 bytes, but unlike records and fields its size field counts the
+      header itself.
+    status: community
+    source: en.uesp.net "Skyrim Mod:Mod File Format", section "Groups" ("in contrast to records and
+      fields, whose sizes do not include their header sizes")
+    tags:
+      - plugin
+      - grup
+      - parser
+      - byte-layout
+      - gotcha
+    detail: Getting this backwards is the classic off-by-24 that makes a whole-file walk desynchronise
+      after the first group.
+  - id: sse.plugins.format.modindex-equal-to-master-count-is-the-plugins-own
+    claim: Inside a plugin, a FormID's top byte (its mod index) counts the plugin's masters in MAST
+      order, so a record whose mod index equals the number of masters is the plugin's own new
+      record, one with a lower index overrides that master's record, and one with a higher index
+      references a master the header does not list. This is how "new record" is decided for the
+      light-plugin range check without any load order.
+    status: community
+    source: en.uesp.net "Skyrim Mod:Mod File Format/TES4", MAST ("The modindex of formids in the mod
+      file match the order of masters given here"); Ortham/esplugin src/record_id.rs
+      `ResolvedRecordId::from_form_id`, which matches a raw FormID against each master's
+      `mod_index_mask` (`i << 24`, built in src/plugin.rs `hashed_masters`) and sets
+      `overridden_record` only on a match, as used by src/plugin.rs `is_valid_as_light_plugin`;
+      dyndolod.info "Internal File FormID Is A HITME"
+    tags:
+      - formid
+      - modindex
+      - masters
+      - new-record
+      - override
+      - validator
+    detail: "esplugin splits the same FormIDs in two rather than three: a mod index that matches no
+      master is not an override, so the higher-than-master-count case counts as the plugin's own for
+      the light-range check. DynDOLOD, on that case - \"HITME stands for HIgher Than Master Entry\",
+      the message coming from xEdit's internal error check, and \"it can not be determined if the
+      record is an overwrite (and which master plugin it belongs to) or if the record is added as a
+      master record by the plugin itself\". ModWright's `sse.plugins.esl-range` warns on it for any
+      plugin it walks, light or not, and range-checks it as esplugin does."

package/knowledge/skyrimse/skse.runtime.yaml

@@ -0,0 +1,283 @@
+game: skyrimse
+topic: skse.runtime
+title: SKSE64 — build-to-runtime coupling, log paths and lines, the Address Library, and the Papyrus
+  runtime log
+facts:
+  - id: sse.skse.runtime.build-is-locked-to-one-runtime
+    claim: An SKSE64 build targets exactly one game runtime. As of 2026-09-11 the Anniversary Edition
+      build is 2.3.1 for game 1.7.104, the GOG build 2.2.6, the Special Edition build 2.0.20 for
+      game 1.5.97 and the VR build 2.0.12 for game 1.4.15.
+    status: community
+    source: skse.silverlock.org, the Downloads section; corroborated by github.com/ianpatt/skse64
+      skse64_common/skse_version.h (CURRENT_RELEASE_SKSE_STR "2.3.1", CURRENT_RELEASE_RUNTIME
+      RUNTIME_VERSION_1_7_104)
+    tags:
+      - skse
+      - version
+      - runtime
+      - compatibility
+      - compat-matrix
+    detail: The same page carries a warning dated 2026-08-14 that Bethesda has announced an incoming
+      Skyrim update, so this pairing is current but explicitly volatile. The GOG build's target
+      runtime is stated inconsistently - the front page says 1.6.1179, both readmes say 1.6.1170,
+      and skse_version.h enumerates both - so a compat rule should accept either.
+  - id: sse.skse.runtime.no-microsoft-store-or-epic-support
+    claim: SKSE does not support the Windows Store / Game Pass release of Skyrim, and does not support
+      the Epic Games Store release.
+    status: community
+    source: skse.silverlock.org, the "Compatibility" section
+    tags:
+      - skse
+      - compatibility
+      - storefront
+      - microsoft-store
+      - epic
+      - compat-matrix
+    detail: Worth a blocking advisory rule, because both storefronts are detectable - MO2 probes for
+      EOSSDK-Win64-Shipping.dll to identify an Epic install, and SKSE's own source has an EPIC
+      save-folder branch even though the build is unsupported.
+  - id: sse.skse.runtime.installs-into-the-game-root
+    claim: SKSE's .dll and .exe files go in the game root beside SkyrimSE.exe, not in Data. Its .pex
+      files go to Data\Scripts\ and are needed by every user; its .psc sources go to
+      Data\Scripts\Source\. The game must be started through skse64_loader.exe.
+    status: community
+    source: github.com/ianpatt/skse64 skse64_readme.txt, section "[ Installation ]"
+    tags:
+      - skse
+      - install
+      - paths
+      - loader
+      - deployment
+    detail: Verbatim - "Do not copy these files to the Data folder as with a normal mod" and "You or
+      your mod manager must continue to use the loader for SKSE64 to work." skse.ini is not shipped;
+      a user creates an empty <skyrim root>\Data\SKSE\skse.ini.
+  - id: sse.skse.runtime.log-directory-is-storefront-dependent
+    claim: SKSE writes skse64_loader.log and skse64.log to Documents\My Games\<SAVE_FOLDER_NAME>\SKSE\,
+      where SAVE_FOLDER_NAME is "Skyrim Special Edition" for Steam, "Skyrim Special Edition GOG" for
+      GOG and "Skyrim Special Edition EPIC" for Epic.
+    status: community
+    source: github.com/ianpatt/skse64 skse64_loader/main.cpp and skse64/skse64.cpp
+      (gLog.OpenRelative(CSIDL_MYDOCUMENTS, "\\My Games\\" SAVE_FOLDER_NAME "\\SKSE\\...")), with
+      SAVE_FOLDER_NAME selected by runtime type in skse64_common/skse_version.h
+    tags:
+      - skse
+      - logs
+      - path
+      - storefront
+      - gog
+      - epic
+    detail: The same three folder names libloadorder and MO2 derive independently for %LOCALAPPDATA% and
+      Documents\My Games. A log source that hardcodes the Steam name finds nothing on a GOG or Epic
+      install.
+  - id: sse.skse.runtime.logs-carry-no-timestamp-or-level
+    claim: Neither skse64.log nor skse64_loader.log carries a per-line timestamp or severity prefix. The
+      _ERROR / _WARNING / _MESSAGE macros only decide whether a line is written, not how it is
+      labelled.
+    status: community
+    source: github.com/ianpatt/common common/IDebugLog.cpp, IDebugLog::Message (writes an empty header
+      plus the message, indented by indentLevel * 4 spaces) and IDebugLog.h's level macros
+    tags:
+      - skse
+      - logs
+      - format
+      - classifier
+      - gotcha
+    detail: A log classifier must match on message text here, not on a bracketed level token - the
+      opposite of the spdlog-shaped logs most other modding frameworks produce.
+  - id: sse.skse.runtime.plugin-load-line-format
+    claim: Every SKSE plugin outcome is one line of the form `plugin <dll> (<dataVersion:08X> <name>
+      <pluginVersion:08X>) <reason> (handle N)`, where a failure inserts a numeric error code
+      between the reason and `(handle N)` and a success has the reason `loaded correctly` with no
+      code.
+    status: community
+    source: github.com/ianpatt/skse64 skse64/PluginManager.cpp, PluginManager::LogPluginLoadError and
+      PluginManager::LoadComplete
+    tags:
+      - skse
+      - logs
+      - format
+      - classifier
+      - plugin
+    detail: The extra numeric field before "(handle N)" is the discriminator between success and
+      failure. SKSE 2.0.20 printed the plugin's full path instead of the DLL name and had no
+      error-code field, so an old-format line also dates the install to the SE branch.
+  - id: sse.skse.runtime.plugin-rejection-reasons
+    claim: SKSE 2.3.1's rejection reasons are a fixed set of literal strings, including "disabled,
+      incompatible with current version of the game", "disabled, address library needs to be
+      updated", "must be recompiled for new address library", "couldn't load plugin", "no version
+      data", "reported as incompatible during load" and "disabled, requires newer script extender".
+    status: community
+    source: github.com/ianpatt/skse64 skse64/PluginManager.cpp (ScanPlugins, CheckAddressLibrary,
+      SafeCallLoadPlugin, CheckPluginCompatibility, InstallPlugins)
+    tags:
+      - skse
+      - logs
+      - classifier
+      - rejection
+      - plugin
+      - version
+    detail: Also present - "LE plugin cannot be used with SE", "does not appear to be an SKSE plugin",
+      "disabled, bad version data", "disabled, no name specified", "disabled, unsupported version
+      independence method", "disabled, only compatible with versions earlier than 1.6.629", and two
+      SEH variants. SKSE 2.0.20 additionally emitted "reported as incompatible during query". Two
+      user-facing remaps exist - skee64.dll is reported as RaceMenu, and EngineFixes.dll failing
+      with ERROR_MOD_NOT_FOUND is reported as "SSE Engine Fixes (Part 2) not installed".
+  - id: sse.skse.runtime.plugin-declares-a-data-symbol-not-a-query-function
+    claim: A modern SKSE plugin exports a data symbol SKSEPluginVersionData SKSEPlugin_Version plus bool
+      SKSEPlugin_Load(const SKSEInterface*). There is no SKSEPlugin_Query in the AE-era API; that
+      was the pre-AE SE model.
+    status: community
+    source: github.com/ianpatt/skse64 skse64/PluginAPI.h, the plugin API docs and the
+      SKSEPluginVersionData struct
+    tags:
+      - skse
+      - plugin
+      - api
+      - version
+      - declaration
+    detail: The struct carries dataVersion, pluginVersion, name[256], author[256], supportEmail[252],
+      versionIndependenceEx, versionIndependence, compatibleVersions[16] (a zero-terminated list of
+      RUNTIME_VERSION_ values) and seVersionRequired. A plugin that is not version-independent must
+      name the exact runtime in compatibleVersions or it is rejected.
+  - id: sse.skse.runtime.address-library-file-naming
+    claim: SKSE looks for the Address Library at
+      Data\SKSE\Plugins\versionlib-<major>-<minor>-<build>-0.bin, with a literal 0 in the fourth
+      position rather than the runtime type.
+    status: community
+    source: github.com/ianpatt/skse64 skse64/PluginManager.cpp, PluginManager::CheckAddressLibrary's
+      _snprintf_s format string
+    tags:
+      - skse
+      - address-library
+      - versionlib
+      - path
+      - compat-matrix
+  - id: sse.skse.runtime.address-library-stem-differs-between-se-and-ae
+    claim: The Address Library filename stem differs by branch, not only by version numbers - AE builds
+      use versionlib-<version>.bin, SE (1.5.97) uses version-<version>.bin, and VR uses
+      version-<version>.csv.
+    status: community
+    source: github.com/alandtse/CommonLibSSE-NG include/REL/IDDB.h, IDDB::load() (selects
+      "versionlib-{}.bin" when Module::IsAE(), "version-{}.bin" otherwise, "version-{}.csv" for VR)
+    tags:
+      - skse
+      - address-library
+      - versionlib
+      - se-vs-ae
+      - vr
+      - compat-matrix
+      - gotcha
+    detail: The same header names four database generations - SSEv1, SSEv2, SSEv5 and VR - with format 5
+      arriving at 1.7.99 and being dense and direct-indexed where formats 1 and 2 are sparse and
+      sorted by id. That transition is what SKSE's "must be recompiled for new address library"
+      rejection and the kVersionIndependentEx_AddressLibraryV5 flag are about.
+  - id: sse.skse.runtime.address-library-failure-strings
+    claim: An Address Library failure surfaces in the plugin's own log or a message box, not in
+      skse64.log, with one of three literal messages - "Unsupported address library format", "Failed
+      to find the id within the address library", or "Failed to locate an appropriate address
+      library with the path".
+    status: community
+    source: github.com/Ryan-rsm-McKenzie/CommonLibSSE include/REL/Relocation.h, IDDatabase's
+      report_and_fail calls
+    tags:
+      - skse
+      - address-library
+      - logs
+      - classifier
+      - error
+  - id: sse.skse.runtime.commonlibsse-plugin-logs
+    claim: A CommonLibSSE-based plugin writes <PluginName>.log into the same SKSE log folder, with the
+      default spdlog pattern "[%Y-%m-%d %H:%M:%S.%e] [%n] [%l] [%t] [%s:%#] %v".
+    status: community
+    source: github.com/alandtse/CrashLoggerSSE src/main.cpp, InitializeLogging() and its comment naming
+      "the standard SKSE logging directory at Documents/My Games/Skyrim Special Edition/SKSE (or
+      Skyrim VR if you are using VR)"
+    tags:
+      - skse
+      - logs
+      - plugin
+      - spdlog
+      - format
+      - classifier
+    detail: So plugin logs do carry a bracketed level, unlike SKSE's own two. Individual plugins may
+      override the pattern.
+  - id: sse.skse.runtime.crashloggersse-log-naming
+    claim: CrashLoggerSSE writes crash-YYYY-MM-DD-HH-MM-SS.log, by default into the SKSE log folder but
+      into a configurable crashDirectory when one is set, with the logger pattern set to "%v" so no
+      line carries a timestamp of its own.
+    status: community
+    source: github.com/alandtse/CrashLoggerSSE src/Crash/CommonHeader.cpp (get_timestamped_log, called
+      with the prefix "crash-") and src/main.cpp (SetupCrashLogPath)
+    tags:
+      - crash-log
+      - crashloggersse
+      - path
+      - format
+      - classifier
+    detail: "Its section headers are PROBABLE CALL STACK:, RAW CALL STACK:, REGISTERS:, STACK:,
+      MODULES:, SKSE PLUGINS: and PLUGINS:, and its light plugin lines read [FE:53C] Name.esp with a
+      colon. Trainwreck, the other common crash logger, uses SCRIPT EXTENDER PLUGINS: and [FE 53C]
+      with a space, which is the cheapest way to tell the two formats apart."
+  - id: sse.skse.runtime.papyrus-log-needs-enabling
+    claim: The Papyrus log is off by default and is enabled by [Papyrus] bEnableLogging=1 in Skyrim.ini,
+      usually alongside bEnableTrace=1 and bLoadDebugInformation=1 (which adds line numbers to error
+      traces at the cost of memory).
+    status: community
+    source: stepmodifications.org "Guide:Skyrim INI/Papyrus" (the [Papyrus] settings table) and
+      "Guide:Troubleshooting", the Clean Savegame Procedure
+    tags:
+      - papyrus
+      - logs
+      - ini
+      - skyrim-ini
+      - enablement
+    detail: "An absent Papyrus log almost always means bEnableLogging is 0, not that anything is broken
+      - which is exactly what a surface's explainMissing hook should say. The same section's
+      defaults are self-validating against a real log's header line \"Memory page: 128 (min) 512
+      (max) 76800 (max total)\"."
+  - id: sse.skse.runtime.papyrus-log-path-and-rotation
+    claim: "With logging on the game writes Documents\\My Games\\Skyrim Special
+      Edition\\Logs\\Script\\Papyrus.#.log for # from 0 to 3, where Papyrus.0.log is always the
+      newest and Papyrus.3.log the oldest."
+    status: community
+    source: stepmodifications.org "Guide:Troubleshooting", section "Clean Savegame Procedure"
+    tags:
+      - papyrus
+      - logs
+      - path
+      - rotation
+    detail: Because 0 is replaced rather than appended on each run, an incremental log reader will
+      legitimately see this file as rotated on every launch. Whether the Logs\Script folder follows
+      the GOG/EPIC rename that SKSE's SAVE_FOLDER_NAME applies is not sourced.
+  - id: sse.skse.runtime.papyrus-line-format
+    claim: 'A Papyrus log line is "[MM/DD/YYYY - HH:MM:SSAM] <message>", with a capitalized "Error: " or
+      "Warning: " immediately after the bracket when present. An error with a stack is followed by a
+      bare "stack:" line and tab-indented frames of the form <receiver>.<script>.<function>() -
+      "<source>" Line <n>.'
+    status: community
+    source: a captured Papyrus.0.log at github.com/sswalters/Skyrim/blob/master/papyrus%20log%200
+    tags:
+      - papyrus
+      - logs
+      - format
+      - classifier
+      - stack-trace
+    detail: <receiver> is one of <unknown self>, [EditorID (FormID)], [alias <Name> on quest <Quest>
+      (FormID)], or [ (FormID)] when the editor ID is blank; <source> is a .psc name, "<native>" or
+      "<savegame>"; "Line ?" appears when bLoadDebugInformation is off.
+  - id: sse.skse.runtime.papyrus-missing-store-line-has-no-severity
+    claim: 'The line `Cannot open store for class "X", missing file?` - meaning a script the save
+      references is no longer installed - carries no Error: or Warning: prefix at all.'
+    status: community
+    source: the same captured Papyrus.0.log (github.com/sswalters/Skyrim/blob/master/papyrus%20log%200)
+    tags:
+      - papyrus
+      - logs
+      - classifier
+      - gotcha
+      - missing-script
+    detail: 'One of the most diagnostic lines in the whole log, and a level-gated classifier will miss
+      every instance of it. The companion families that DO carry a prefix are "Error: Unable to bind
+      script X to Y (FormID) because their base types do not match" and "Warning: Property X on
+      script Y attached to Z (FormID) cannot be initialized because the script no longer contains
+      that property".'