This file records third-party attribution for code and assets bundled with AgentOS. It covers:
- Core runtime modules derived from OpenSquilla (Apache-2.0); see the
first section below. Releases up to and including the v2026.7.19 release tag additionally
shipped the V4 Phase 3 router engine module
(
src/agentos/agentos_router/v4_phase3.py) and the V4 Phase 3 model bundle (OpenSquilla-trained weights and inference code); both have since been removed from the tree — see the first section below for the historical record. - The bundled skill descriptors under
src/agentos/skills/bundled/, which include OpenClaw-derived MIT descriptors and AgentOS-original descriptors. - The bundled pptx skill references the python-pptx and PptxGenJS libraries; AgentOS does not vendor those libraries, but the skill instructs the agent runtime to invoke them and is documented here for transparency.
- The bundled BGE (bge-small-zh-v1.5) ONNX export under
src/agentos/memory/models/bge_onnx/, used by local memory embedding (historically also by the V4 Phase 3 router's BGE feature channel). - The bundled all-MiniLM-L6-v2 INT8 ONNX export under
src/agentos/memory/models/embeddings/all-MiniLM-L6-v2-int8/, used by the Pilot router's feature builder. - The checked-in Pilot router golden evaluation set under
tests/test_agentos_router/data/pilot_golden.jsonl, whose rows are derived from the WildChat-1M dataset (ODC-BY 1.0). - The built-in tokenjuice tool-result projection backend and bundled
reduction rules under
src/agentos/plugins/tokenjuice/. - The cron prompt-injection scanner was reviewed against Hermes Agent reference material; the MIT notice is reproduced below for conservative attribution.
- The production dependencies and font assets bundled into the React Control
UI. Exact package versions and complete upstream license texts are generated
from
frontend/package-lock.jsonplus the checked-in font licenses into the shippedstatic/dist/THIRD_PARTY_LICENSES.txt.
- Component: core runtime modules under
src/agentos/, and (in releases up to and including the v2026.7.19 release tag) the V4 Phase 3 local ML router bundle (trained model weights and inference code) undersrc/agentos/agentos_router/models/v4.2_phase3_inference/. - Upstream project: https://github.com/opensquilla/opensquilla
- License: Apache License 2.0
- Copyright notice: OpenSquilla contributors (the upstream project ships the stock Apache-2.0 text without a filled-in copyright line and no NOTICE file).
AgentOS is built on OpenSquilla. Parts of the AgentOS core were copied from and then substantially modified relative to the upstream project. The highest-overlap modules include:
src/agentos/application/approval_queue.pyandsrc/agentos/gateway/approval_queue.py— approval queue handlingsrc/agentos/cli/agent_cmd.py— agent CLI command surfacesrc/agentos/channels/command_registry.py— channel slash-command dispatchsrc/agentos/gateway_client.pyandsrc/agentos/cli/gateway_client.py— gateway client plumbingsrc/agentos/agentos_router/v4_phase3.py— router phase logic (shipped in releases up to and including the v2026.7.19 release tag; since removed from the tree with the rest of the V4 Phase 3 router). Its tier-resolution helper survives in modified form assrc/agentos/agentos_router/tiers_util.py, which remains OpenSquilla-derived.
Releases up to and including the v2026.7.19 release tag bundled the local ML router assets
under src/agentos/agentos_router/models/v4.2_phase3_inference/. That bundle
is OpenSquilla's, carried over from upstream
src/opensquilla/squilla_router/models/v4.2_phase3_inference/. It is not
trained or authored by the AgentOS contributors. The bundle and the
OpenSquilla-derived engine module src/agentos/agentos_router/v4_phase3.py
have since been removed from the tree and no longer ship in the wheel (the
default router strategy is the AgentOS-trained pilot-v1). This notice is
retained for the releases that shipped them, which covered:
lgbm_main.binandlgbm_aux.bin— LightGBM boosters for the router heads.mlp/model.onnxandmlp/scaler.joblib— the PyTorch-exported MLP head and its scaler.features/tfidf.pkl,features/svd.pkl,features/config.pkl, andfeatures/bge_pca.joblib— fitted scikit-learn/joblib feature artifacts.runtime_src/src/router/**— the inference core the router loads at runtime.router.runtime.yaml,version.json,inference_manifest.json— runtime configuration and inference metadata.
The weights were used byte-for-byte unmodified: lgbm_main.bin carried the
same Git LFS object as upstream
(sha256:5f312db09577bbaf30f87358941974eef6edce7f1424d0e9de21cbd38a646d53,
39684725 bytes). Modifications made by the AgentOS contributors were limited to
namespace/branding renames, and to runtime_src/.../inference/artifacts.py,
which resolved the BGE export from the shared
src/agentos/memory/models/bge_onnx/ location so it shipped once instead of
twice. The bundle's PROVENANCE.md (present in the releases that shipped it,
and in this repository's git history) records the per-file detail.
Note that only the BGE embedding channel was third-party relative to OpenSquilla (MIT; see the BAAI section below). The routing decision itself came from OpenSquilla's own trained LightGBM and MLP heads.
Other modules across the runtime may also contain OpenSquilla-derived
code in modified form. In accordance with Section 4(b) of the Apache
License 2.0, this notice records that the derived files have been
modified by the AgentOS contributors. The entire AgentOS repository is
licensed under the Apache License 2.0 (see LICENSE), so the upstream
license terms apply uniformly; the full license text is included in the
LICENSE file at the repository root.
- Component: SKILL.md frontmatter and instruction text for these bundled skills:
sub-agentcrongithubnano-pdfsummarizetmuxweather
- Upstream project: https://github.com/openclaw/openclaw
- License: MIT
- Copyright notice: Copyright (c) 2025 Peter Steinberger
Note: the sub-agent descriptor retains OpenClaw upstream lineage and MIT
attribution.
The descriptor text instructs the agent runtime how to use built-in skill surfaces and external tools; AgentOS does not redistribute third-party CLIs through these descriptors. Per the MIT license, the upstream copyright and permission notice are reproduced below in their entirety and apply to the OpenClaw-derived bundled descriptor files.
MIT License
Copyright (c) 2025 Peter Steinberger
Permission is hereby granted, free of charge, to any person obtaining a copy
of this software and associated documentation files (the "Software"), to deal
in the Software without restriction, including without limitation the rights
to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
copies of the Software, and to permit persons to whom the Software is
furnished to do so, subject to the following conditions:
The above copyright notice and this permission notice shall be included in all
copies or substantial portions of the Software.
THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
SOFTWARE.
- Component: SKILL.md instruction text for these bundled skills, plus
gmgn-holder-analysis/scripts/analyze.py,gmgn-wallet-analysis/scripts/analyze.py, andgmgn-wallet-score/scripts/score.py:gmgn-cookinggmgn-holder-analysisgmgn-marketgmgn-portfoliogmgn-swapgmgn-tokengmgn-trackgmgn-wallet-analysisgmgn-wallet-score
- Upstream project: https://github.com/GMGNAI/gmgn-skills
- License: MIT
- Copyright notice: Copyright (c) 2025 GMGN
These descriptors drive the third-party gmgn-cli npm package, which AgentOS
does not redistribute — an operator installs it themselves and supplies
their own GMGN_API_KEY. Only the descriptor text and the helper scripts are
vendored here; AgentOS added the provenance and metadata.agentos frontmatter
blocks, re-pointed the helper-script paths at {baseDir}, folded the
gmgn-wallet-score description into a YAML block scalar so its frontmatter
parses, and lifted that skill's upstream inline analyzer into
scripts/score.py (argv in place of the <FILL_IN_*> placeholders), matching
what was already done for gmgn-holder-analysis. The two wallet skills'
description fields were also shortened to the length of their bundled
siblings; upstream's run roughly twice as long and did not fit the shipped
skills-prompt budget. Per the MIT license, the upstream copyright and permission notice
are reproduced below in their entirety and apply to the GMGN-derived bundled
files.
MIT License
Copyright (c) 2025 GMGN
Permission is hereby granted, free of charge, to any person obtaining a copy
of this software and associated documentation files (the "Software"), to deal
in the Software without restriction, including without limitation the rights
to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
copies of the Software, and to permit persons to whom the Software is
furnished to do so, subject to the following conditions:
The above copyright notice and this permission notice shall be included in all
copies or substantial portions of the Software.
THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
SOFTWARE.
These bundled skill descriptors are authored and maintained by AgentOS and
are released under AgentOS's repository license (Apache-2.0; see LICENSE):
agentoscroncron-watchersdeep-researchdocxgit-diffgithubhistory-explorerhtml-to-pdfhttp-fetchmemorymulti-search-enginemusebooknano-pdfpdf-toolkitpoolsdotfun-token-launcherpptxrobinhood-agentic-tradingrobinhood-chain-stocksrobinhood-rwa-addressessenior-unilp-managerstack-trace-generic-probestack-trace-go-probestack-trace-js-probestack-trace-python-probestack-trace-rust-probesub-agentsrt-from-scriptsubtitle-burnersummarizetext-file-readtitle-card-imagetmuxvideo-still-animatortoken-burnerwallet-tradingweatherxlsxadvanced-dubbing-studiomusic-and-singing-studiovoice-clone-labvoice-conversion-studiovoiceover-studio
The musebook skill descriptor and its scripts/muse.py are AgentOS-original.
references/muse.txt is a verbatim copy of the onboarding musebook.lol
publishes for agents (https://musebook.lol/muse.txt), retrieved 2026-09-17. It
is redistributed unmodified as the authoritative description of a public wire
protocol and carries no license header of its own.
- Component: built-in tokenjuice tool-result projection backend and bundled
reduction rules under
src/agentos/plugins/tokenjuice/. - Upstream project: https://github.com/vincentkoc/tokenjuice
- License: MIT
- Copyright notice: Copyright (c) 2026 Vincent Koc
AgentOS includes a Python adaptation of tokenjuice's rule-driven reducer
and bundles reduction rules derived from the upstream project. AgentOS does
not depend on the upstream tokenjuice npm package at runtime. Additional
provenance is recorded in
src/agentos/plugins/tokenjuice/PROVENANCE.md; the MIT license text is
also shipped with that package as LICENSE.tokenjuice.
MIT License
Copyright (c) 2026 Vincent Koc
Permission is hereby granted, free of charge, to any person obtaining a copy
of this software and associated documentation files (the "Software"), to deal
in the Software without restriction, including without limitation the rights
to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
copies of the Software, and to permit persons to whom the Software is
furnished to do so, subject to the following conditions:
The above copyright notice and this permission notice shall be included in all
copies or substantial portions of the Software.
THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
SOFTWARE.
- Components:
src/agentos/tools/builtin/x_search.py— adapted fromtools/x_search_tool.py. The tool's request shape, client-side date validation, citation extraction, and degraded-result semantics follow the upstream implementation; the HTTP client, retry budgeting, and AgentOS wiring are original.src/agentos/xai_oauth.py— adapted from the xAI OAuth section ofhermes_cli/auth.py. The device-code flow, OIDC discovery, endpoint and inference-origin pinning, refresh-skew heuristics, tier-denial (HTTP 403) handling, and dead-token quarantine follow the upstream implementation; the token store, async refresh path, and network-free availability check are original.src/agentos/tools/browser_supervisor.py— adapted fromtools/browser_supervisor.py. The CDP dialog-interception model (must_respond / auto_dismiss / auto_accept), the pending-dialog state machine, console capture, and the per-session supervisor registry follow the upstream implementation; the pluggable transport, WebSocket loop, and AgentOS wiring are original.src/agentos/tools/browser_eval_policy.py— adapted from the eval-policy helpers intools/browser_tool.py(_enforce_browser_eval_policy,_risky_browser_eval_reason,_sensitive_browser_eval_token_reason,_expression_targets_private_url,_redact_browser_output). The risky- primitive denylist, string-literal deobfuscation, URL-literal SSRF pre-scan, and output redaction follow the upstream implementation.src/agentos/tools/agent_browser.pyandsrc/agentos/tools/builtin/browser.py— original AgentOS code that reuses the upstream lessons on subprocess env-scrubbing (GHSA-m4m8-xjp4-5rmm), CDP-URL redaction, first-open timeouts, and the session model; the engine invoked (agent-browser) is the same one Hermes uses.- Cron prompt-injection scanner reference material (reviewed, not copied).
- Upstream project: https://github.com/NousResearch/hermes-agent
- License: MIT
- Copyright notice: Copyright (c) 2025 Nous Research
AgentOS does not redistribute Hermes Agent as a whole.
MIT License
Copyright (c) 2025 Nous Research
Permission is hereby granted, free of charge, to any person obtaining a copy
of this software and associated documentation files (the "Software"), to deal
in the Software without restriction, including without limitation the rights
to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
copies of the Software, and to permit persons to whom the Software is
furnished to do so, subject to the following conditions:
The above copyright notice and this permission notice shall be included in all
copies or substantial portions of the Software.
THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
SOFTWARE.
- Component: SKILL.md frontmatter and instruction text for these bundled skills:
ai-video-scriptdeep-researchdocxhtml-to-pdfmulti-search-enginenano-banana-propdf-toolkitpptxseedance-2-promptvideo-mergerxlsx
- Upstream registry: https://clawhub.ai
- License: MIT-0 (Public-domain-equivalent; no attribution required, but
each skill records its specific upstream slug in its own
THIRD_PARTY_NOTICES.mdfor transparency)
These bundled skills record their ClawHub source slug in SKILL.md frontmatter
and, when present, the skill-local THIRD_PARTY_NOTICES.md. ClawHub's MIT-0
default license permits unlimited use, modification, and redistribution without
attribution.
MIT No Attribution
Copyright <YEAR> <COPYRIGHT HOLDER>
Permission is hereby granted, free of charge, to any person obtaining a copy
of this software and associated documentation files (the "Software"), to deal
in the Software without restriction, including without limitation the rights
to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
copies of the Software, and to permit persons to whom the Software is
furnished to do so.
THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
SOFTWARE.
- Component: BAAI/bge-small-zh-v1.5 embedding model and tokenizer assets.
- Upstream model: https://huggingface.co/BAAI/bge-small-zh-v1.5
- Upstream project: https://github.com/FlagOpen/FlagEmbedding
- License: MIT
- Copyright notice: Copyright (c) 2022 staoxiao
The bundled local memory embedding assets contain an ONNX export and tokenizer files derived from the BAAI bge-small-zh-v1.5 model. The upstream Hugging Face model card marks the model as MIT licensed and states that the released models can be used for commercial purposes free of charge.
MIT License
Copyright (c) 2022 staoxiao
Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions:
The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software.
THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
- Component: sentence-transformers/all-MiniLM-L6-v2 embedding model and
tokenizer assets, bundled as an INT8 ONNX export under
src/agentos/memory/models/embeddings/all-MiniLM-L6-v2-int8/, used by the Pilot router's feature builder. - Upstream model: https://huggingface.co/sentence-transformers/all-MiniLM-L6-v2
- Upstream project: https://github.com/UKPLab/sentence-transformers
- License: Apache License 2.0
- Copyright notice: sentence-transformers contributors (the upstream model card and library ship under the stock Apache-2.0 text without a filled-in copyright line).
The bundled export is an INT8 dynamic quantization (onnxruntime,
avx512_vnni, per-channel) of the upstream FP32 ONNX weights, produced by
scripts/pilot_router/export_embedder.py. It carries forward the exact
upstream Hugging Face revision and tokenizer recorded in the export's own
export_meta.json:
- HF revision:
1110a243fdf4706b3f48f1d95db1a4f5529b4d41 - Tokenizer (
tokenizer.json) sha256:da0e79933b9ed51798a3ae27893d3c5fa4a201126cef75586296df9b4d2c62a0
The entire AgentOS repository is licensed under the Apache License 2.0 (see
LICENSE), so the upstream license terms apply uniformly; the full license
text is included in the LICENSE file at the repository root.
- Component: the Pilot router's checked-in golden evaluation set,
tests/test_agentos_router/data/pilot_golden.jsonl. Its rows are derived from the WildChat-1M dataset. - Upstream dataset: https://huggingface.co/datasets/allenai/WildChat-1M
- Publisher: Allen Institute for AI (AI2)
- License: ODC-BY 1.0 (Open Data Commons Attribution License)
- License text: https://opendatacommons.org/licenses/by/1-0/
The WildChat-1M dataset is released by AI2 under the Open Data Commons
Attribution License (ODC-BY) 1.0. The Pilot router training corpus is sampled
from WildChat-1M, and the corpus rows themselves are never committed to this
repository (they live only under the git-ignored scripts/pilot_router/data/;
see scripts/pilot_router/DATA.md for the full corpus provenance and license
gate). The one WildChat-derived data file that is checked in is the golden
evaluation set tests/test_agentos_router/data/pilot_golden.jsonl — a small set
of user-turn rows drawn from WildChat and used as a routing-quality regression
fixture.
ODC-BY 1.0 is a permissive open-data license whose substantive obligation is attribution: any public use of the database or a work produced from it must keep the attribution notice and the ODC-BY notice intact. It imposes no share-alike / copyleft requirement and no restriction on commercial use. This notice, together with the ODC-BY reference above, provides that attribution for the checked-in WildChat-derived golden-set rows.
The browser bundle is built from the lockfile-pinned packages below. The
release build generates static/dist/THIRD_PARTY_LICENSES.txt from the
installed package metadata, upstream package license files, and the checked-in
font licenses. That generated file is included in every wheel and source
distribution alongside the minified browser assets, so the exact copyright
and permission notices shipped by each upstream project remain available even
when multiple packages use the same license.
@radix-ui/react-compose-refs@1.1.3— MIT@radix-ui/react-slot@1.3.0— MIT@tabler/icons@3.45.0— MIT@tabler/icons-react@3.45.0— MIT@tanstack/query-core@5.101.2— MIT@tanstack/react-query@5.101.2— MITclass-variance-authority@0.7.1— Apache-2.0clsx@2.1.1— MITcookie@1.1.1— MITdompurify@3.4.12— MPL-2.0 or Apache-2.0fancy-canvas@2.1.0— MITframer-motion@12.42.2— MIThighlight.js@11.11.1— BSD-3-Clauselightweight-charts@5.2.0— Apache-2.0lucide-react@1.25.0— ISCmarked@18.0.7— MITmotion@12.42.2— MITmotion-dom@12.42.2— MITmotion-utils@12.39.0— MITqrcode-generator@1.5.2— MITreact@19.2.7— MITreact-dom@19.2.7— MITreact-router@7.18.1— MITscheduler@0.27.0— MITset-cookie-parser@2.7.2— MITsonner@2.0.7— MITtailwind-merge@3.6.0— MITtslib@2.8.1— 0BSDzustand@5.0.14— MIT
The type-only @types/* and csstype packages installed to compile the
TypeScript source do not enter the browser bundle and are intentionally
omitted from this production list.
fancy-canvas@2.1.0 (a dependency of lightweight-charts) publishes a files
allowlist covering only its compiled JavaScript and type declarations, so its
MIT text is absent from the npm tarball. The upstream license from
https://github.com/tradingview/fancy-canvas is kept verbatim at
frontend/vendor-licenses/fancy-canvas-LICENSE.txt, and the Control UI builder
appends it to the generated ledger marked as vendored.
qrcode-generator@1.5.2 (the Control UI's local QR renderer) likewise
publishes only its sources, README and type declarations. The MIT text from
https://github.com/kazuhikoarase/qrcode-generator (Copyright (c) 2009 Kazuhiko
Arase) is kept verbatim at
frontend/vendor-licenses/qrcode-generator-LICENSE.txt and appended the same
way. Packages that ship no license and have no vendored copy still fail the
build.
- Inter Variable — Copyright (c) 2016 The Inter Project Authors (https://github.com/rsms/inter), SIL Open Font License 1.1.
- JetBrains Mono Variable — Copyright 2020 The JetBrains Mono Project Authors (https://github.com/JetBrains/JetBrainsMono), SIL Open Font License 1.1.
The complete upstream notices and OFL 1.1 texts are kept in
frontend/src/assets/fonts/Inter-LICENSE.txt and
frontend/src/assets/fonts/JetBrainsMono-LICENSE.txt. The shared Control UI
builder appends both files verbatim to the license ledger shipped beside the
font binaries.