fux-engine 0.1.0__tar.gz

fux_engine-0.1.0/LICENSE

@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2026 arpit arya
+
+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.

fux_engine-0.1.0/PKG-INFO

@@ -0,0 +1,186 @@
+Metadata-Version: 2.4
+Name: fux-engine
+Version: 0.1.0
+Summary: Fux — a portable, Claude-aware knowledge engine: rules, memory, narrative, and graph in one frontmatter substrate.
+Author-email: arpit arya <arpitarya.me@gmail.com>
+License: MIT
+Project-URL: Homepage, https://github.com/arpitarya/fux
+Project-URL: Repository, https://github.com/arpitarya/fux
+Project-URL: Documentation, https://github.com/arpitarya/fux/blob/main/docs/guide.md
+Project-URL: Bug Tracker, https://github.com/arpitarya/fux/issues
+Keywords: knowledge,rules,graph,claude,frontmatter,llm,memory,mcp
+Classifier: Development Status :: 4 - Beta
+Classifier: Intended Audience :: Developers
+Classifier: License :: OSI Approved :: MIT License
+Classifier: Operating System :: OS Independent
+Classifier: Programming Language :: Python :: 3
+Classifier: Programming Language :: Python :: 3.11
+Classifier: Programming Language :: Python :: 3.12
+Classifier: Programming Language :: Python :: 3.13
+Classifier: Programming Language :: Python :: 3.14
+Classifier: Topic :: Software Development :: Documentation
+Classifier: Topic :: Software Development :: Quality Assurance
+Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
+Requires-Python: >=3.11
+Description-Content-Type: text/markdown
+License-File: LICENSE
+Provides-Extra: embeddings
+Requires-Dist: sentence-transformers>=2.2; extra == "embeddings"
+Provides-Extra: ast
+Requires-Dist: tree-sitter>=0.23; extra == "ast"
+Requires-Dist: tree-sitter-language-pack>=0.7; extra == "ast"
+Provides-Extra: pdf
+Requires-Dist: pypdf>=4.0; extra == "pdf"
+Provides-Extra: dev
+Requires-Dist: pytest>=8; extra == "dev"
+Requires-Dist: build>=1.0; extra == "dev"
+Requires-Dist: twine>=5.0; extra == "dev"
+Dynamic: license-file
+
+# Fux
+
+> A portable, Claude-aware **knowledge engine**. One frontmatter substrate →
+> derived **index**, **graph**, and **memory** views. `$0` deterministic
+> maintenance — no mandatory LLM calls. Continuously referenced, cheaply
+> maintained.
+
+Named after *Johann Joseph Fux*, author of *Gradus ad Parnassum* (1725) — the
+counterpoint treatise every composer learned the rules from. A tool that codifies
+and enforces rules, named after the man who wrote *the* rulebook. Sits beside
+`wagner`, `bach`, `orff`.
+
+Fux **unifies and replaces** three things a project usually runs separately — the
+structural graph (graphify), cross-session memory, and the narrative docs — and
+adds the **business-rules layer** none of them held. See [docs/fux-plan.md](docs/fux-plan.md).
+
+## In plain words (the 5-year-old version)
+
+Think of your code as a big LEGO city. The *reasons* things are built a certain
+way — why the bridge is red, why you add the towers before the walls — usually
+live only in one builder's head. When they go home, nobody remembers why.
+
+**Fux is a notebook for those reasons.** You write each important rule down once —
+*what* it is and *why* — and Fux sticks a tiny one-line list on the cover so your
+helper (Claude) can flip straight to the right page instead of searching the whole
+city every time. Fux also **draws a map** showing which notes belong to which
+buildings, and **checks** that the notes still match the city — and it does all of
+that for free, without phoning anyone for help.
+
+So: write the *why* down once → it's found fast, stays correct, and never gets lost.
+
+## Why
+
+The *why* behind a formula — why current value not invested cost, why
+INR-normalize first, which cost-basis method — usually lives only as an inline
+comment, invisible until someone greps for it. Fux makes that knowledge
+**first-class**: one entry, authored once, served back through a tiny index (read
+first) plus lazily-opened rules (read only when relevant). Lookups run ~5–10×
+cheaper and more correct on every later session — and you don't have to take that
+on faith: **`fux savings`** measures the multiplier from your own file sizes.
+
+## Install
+
+```bash
+git clone git@github.com:arpitarya/fux.git && cd fux
+./install.sh          # → ~/.claude/fux/{engine,global,packs,hooks} + /fux skill
+fux --version
+```
+
+Requires Python ≥ 3.11 (stdlib only — zero third-party dependencies).
+
+## Use
+
+```bash
+cd your-project
+fux init                       # scaffold .fux/ + wire 3 hooks + CLAUDE.md pointer
+fux new formula day-pnl        # scaffold a rule; fill **Rule:/Why:/Edge cases:**
+fux build                      # regenerate INDEX.md + rules.json + graph   ($0)
+fux check --fix                # validate; repair mechanical drift           ($0)
+fux why day-pnl [--history]    # explain a rule (+ how its *why* evolved, via git)
+fux refs src/aggregator.py     # which rules govern this file
+fux recall "how is day P&L computed" --hybrid  # BM25F; RRF-fuse lexical+semantic+graph
+fux seal --all                 # bind rules to an AST fingerprint of their code
+fux coverage                   # % of important files with a governing rule
+fux verify --fuzz              # run invariant `check:`; boundary-fuzz for div-by-zero
+fux mine                       # surface candidate rules latent in the code (drafts)
+fux savings "how is day P&L computed"  # measured token-cost win (+ cumulative ledger)
+fux lint                       # rule *quality*: missing why / code_refs / edges
+fux stats                      # knowledge-health dashboard + score
+fux gate --install             # wire a git pre-commit enforcement hook
+fux mcp                        # serve the substrate to agents over MCP (stdio)
+fux capture                    # queue this session's changes for `fux distill`
+fux serve                      # local dashboard over the generated views
+fux import docs/               # migrate existing markdown → narrative entries
+fux parity                     # is it safe to retire the old graph/docs/memory?
+fux tour                       # ordered ONBOARDING.md
+```
+
+**Complete, example-driven guide to everything Fux does:
+[docs/guide.md](docs/guide.md).** Full command reference: [docs/cli.md](docs/cli.md).
+Authoring a rule: [docs/rule.guide.md](docs/rule.guide.md). Writing a spec:
+[docs/spec.guide.md](docs/spec.guide.md).
+
+## How it works
+
+```
+Tier 0  CLAUDE.md pointer ........ 1 line, always in context
+Tier 1  .fux/out/INDEX.md ........ ~1 line/rule, read FIRST        ← cheap
+Tier 2  .fux/rules/<id>.md ....... opened ONLY when relevant       ← lazy
+Tier 3  .fux/out/{rules,graph}.json  machine lookup + browsing
+```
+
+You maintain **only** the source frontmatter. INDEX, `rules.json`, and the
+interactive `graph.html` regenerate on `fux build`. Three hooks keep it live:
+SessionStart injects the INDEX, PostToolUse reminds you when an edited file's rule
+drifted, Stop validates before the turn ends.
+
+The **graph** merges your rules with code symbols and call edges extracted across
+**Python** (via the stdlib `ast`) and **JS/TS, Go, and Rust** (a brace-matched
+heuristic by default, or **real tree-sitter ASTs** with the optional
+`pip install fux-engine[ast]` extra — same schema, more accuracy, still $0 by
+default), now including **cross-file** `calls` (symbol→symbol) — one navigable
+map of *which rule governs which code*, with community clustering, **PageRank
+centrality** (architectural chokepoints, not just raw degree), and a
+`GRAPH_REPORT.md`. The interactive `graph.html` is built for review *and* agents:
+node/edge-type filters, colour-by (type/community/layer/degree), focus +
+neighbour highlighting, a details panel, and one-click **markdown export** of a
+node's neighbourhood or the visible sub-graph.
+
+`fux recall` is lexical **BM25F** (`$0`) by default, with opt-in **query expansion**
+(glossary synonyms + 1-hop graph neighbours), an opt-in **local** re-rank, and an
+opt-in **RRF hybrid** that fuses lexical ⊕ local-semantic ⊕ graph proximity (no
+API); `fux verify [--fuzz]` runs a rule's invariant `check:` and worked `examples:`,
+optionally boundary-fuzzing for unguarded div-by-zero. **Proof-carrying rules:**
+`fux seal` binds a rule to a normalized-AST fingerprint of its code, so `fux check`
+flags `unsealed` when the governed code changes *structure* (not just its mtime).
+Beyond authoring, Fux **enforces and reports**: `fux lint` grades rule quality,
+`fux stats` scores knowledge health, `fux gate` blocks drift at commit/CI time,
+and `fux mcp` exposes the whole substrate to agents over MCP.
+
+For cross-session memory it stays **authored, not captured**: an opt-in `capture`
+hook queues *which* files changed for `fux distill` (human-confirmed) rather than
+auto-summarising, and `type: memory` entries **decay** after a TTL so stale notes
+stop costing context — with opt-in **usage-weighted decay** (a memory still being
+recalled stays alive; an unused one decays). Every path `$0` and deterministic.
+
+### Layered rules (maintain once, inherit everywhere)
+
+```
+effective ruleset = ~/.claude/fux/global/   (cross-project best practices)
+                  ⊕ ~/.claude/fux/packs/*    (opt-in shareable domain packs)
+                  ⊕ ./.fux/rules/            (this project's domain rules)
+```
+
+`project` overrides `pack` overrides `global`. `fux check` flags conflicts
+instead of silently shadowing.
+
+## Guarantee
+
+Every maintenance command is shell/AST/parse — **no LLM calls**. The only paths
+that call the LLM are the `plan` / `adr` skills, and they ride the session you are
+already in (no background spend). The same "$0, deterministic" promise that made
+graphify trustworthy.
+
+## License
+
+MIT — see [LICENSE](LICENSE).

fux_engine-0.1.0/README.md

@@ -0,0 +1,147 @@
+# Fux
+
+> A portable, Claude-aware **knowledge engine**. One frontmatter substrate →
+> derived **index**, **graph**, and **memory** views. `$0` deterministic
+> maintenance — no mandatory LLM calls. Continuously referenced, cheaply
+> maintained.
+
+Named after *Johann Joseph Fux*, author of *Gradus ad Parnassum* (1725) — the
+counterpoint treatise every composer learned the rules from. A tool that codifies
+and enforces rules, named after the man who wrote *the* rulebook. Sits beside
+`wagner`, `bach`, `orff`.
+
+Fux **unifies and replaces** three things a project usually runs separately — the
+structural graph (graphify), cross-session memory, and the narrative docs — and
+adds the **business-rules layer** none of them held. See [docs/fux-plan.md](docs/fux-plan.md).
+
+## In plain words (the 5-year-old version)
+
+Think of your code as a big LEGO city. The *reasons* things are built a certain
+way — why the bridge is red, why you add the towers before the walls — usually
+live only in one builder's head. When they go home, nobody remembers why.
+
+**Fux is a notebook for those reasons.** You write each important rule down once —
+*what* it is and *why* — and Fux sticks a tiny one-line list on the cover so your
+helper (Claude) can flip straight to the right page instead of searching the whole
+city every time. Fux also **draws a map** showing which notes belong to which
+buildings, and **checks** that the notes still match the city — and it does all of
+that for free, without phoning anyone for help.
+
+So: write the *why* down once → it's found fast, stays correct, and never gets lost.
+
+## Why
+
+The *why* behind a formula — why current value not invested cost, why
+INR-normalize first, which cost-basis method — usually lives only as an inline
+comment, invisible until someone greps for it. Fux makes that knowledge
+**first-class**: one entry, authored once, served back through a tiny index (read
+first) plus lazily-opened rules (read only when relevant). Lookups run ~5–10×
+cheaper and more correct on every later session — and you don't have to take that
+on faith: **`fux savings`** measures the multiplier from your own file sizes.
+
+## Install
+
+```bash
+git clone git@github.com:arpitarya/fux.git && cd fux
+./install.sh          # → ~/.claude/fux/{engine,global,packs,hooks} + /fux skill
+fux --version
+```
+
+Requires Python ≥ 3.11 (stdlib only — zero third-party dependencies).
+
+## Use
+
+```bash
+cd your-project
+fux init                       # scaffold .fux/ + wire 3 hooks + CLAUDE.md pointer
+fux new formula day-pnl        # scaffold a rule; fill **Rule:/Why:/Edge cases:**
+fux build                      # regenerate INDEX.md + rules.json + graph   ($0)
+fux check --fix                # validate; repair mechanical drift           ($0)
+fux why day-pnl [--history]    # explain a rule (+ how its *why* evolved, via git)
+fux refs src/aggregator.py     # which rules govern this file
+fux recall "how is day P&L computed" --hybrid  # BM25F; RRF-fuse lexical+semantic+graph
+fux seal --all                 # bind rules to an AST fingerprint of their code
+fux coverage                   # % of important files with a governing rule
+fux verify --fuzz              # run invariant `check:`; boundary-fuzz for div-by-zero
+fux mine                       # surface candidate rules latent in the code (drafts)
+fux savings "how is day P&L computed"  # measured token-cost win (+ cumulative ledger)
+fux lint                       # rule *quality*: missing why / code_refs / edges
+fux stats                      # knowledge-health dashboard + score
+fux gate --install             # wire a git pre-commit enforcement hook
+fux mcp                        # serve the substrate to agents over MCP (stdio)
+fux capture                    # queue this session's changes for `fux distill`
+fux serve                      # local dashboard over the generated views
+fux import docs/               # migrate existing markdown → narrative entries
+fux parity                     # is it safe to retire the old graph/docs/memory?
+fux tour                       # ordered ONBOARDING.md
+```
+
+**Complete, example-driven guide to everything Fux does:
+[docs/guide.md](docs/guide.md).** Full command reference: [docs/cli.md](docs/cli.md).
+Authoring a rule: [docs/rule.guide.md](docs/rule.guide.md). Writing a spec:
+[docs/spec.guide.md](docs/spec.guide.md).
+
+## How it works
+
+```
+Tier 0  CLAUDE.md pointer ........ 1 line, always in context
+Tier 1  .fux/out/INDEX.md ........ ~1 line/rule, read FIRST        ← cheap
+Tier 2  .fux/rules/<id>.md ....... opened ONLY when relevant       ← lazy
+Tier 3  .fux/out/{rules,graph}.json  machine lookup + browsing
+```
+
+You maintain **only** the source frontmatter. INDEX, `rules.json`, and the
+interactive `graph.html` regenerate on `fux build`. Three hooks keep it live:
+SessionStart injects the INDEX, PostToolUse reminds you when an edited file's rule
+drifted, Stop validates before the turn ends.
+
+The **graph** merges your rules with code symbols and call edges extracted across
+**Python** (via the stdlib `ast`) and **JS/TS, Go, and Rust** (a brace-matched
+heuristic by default, or **real tree-sitter ASTs** with the optional
+`pip install fux-engine[ast]` extra — same schema, more accuracy, still $0 by
+default), now including **cross-file** `calls` (symbol→symbol) — one navigable
+map of *which rule governs which code*, with community clustering, **PageRank
+centrality** (architectural chokepoints, not just raw degree), and a
+`GRAPH_REPORT.md`. The interactive `graph.html` is built for review *and* agents:
+node/edge-type filters, colour-by (type/community/layer/degree), focus +
+neighbour highlighting, a details panel, and one-click **markdown export** of a
+node's neighbourhood or the visible sub-graph.
+
+`fux recall` is lexical **BM25F** (`$0`) by default, with opt-in **query expansion**
+(glossary synonyms + 1-hop graph neighbours), an opt-in **local** re-rank, and an
+opt-in **RRF hybrid** that fuses lexical ⊕ local-semantic ⊕ graph proximity (no
+API); `fux verify [--fuzz]` runs a rule's invariant `check:` and worked `examples:`,
+optionally boundary-fuzzing for unguarded div-by-zero. **Proof-carrying rules:**
+`fux seal` binds a rule to a normalized-AST fingerprint of its code, so `fux check`
+flags `unsealed` when the governed code changes *structure* (not just its mtime).
+Beyond authoring, Fux **enforces and reports**: `fux lint` grades rule quality,
+`fux stats` scores knowledge health, `fux gate` blocks drift at commit/CI time,
+and `fux mcp` exposes the whole substrate to agents over MCP.
+
+For cross-session memory it stays **authored, not captured**: an opt-in `capture`
+hook queues *which* files changed for `fux distill` (human-confirmed) rather than
+auto-summarising, and `type: memory` entries **decay** after a TTL so stale notes
+stop costing context — with opt-in **usage-weighted decay** (a memory still being
+recalled stays alive; an unused one decays). Every path `$0` and deterministic.
+
+### Layered rules (maintain once, inherit everywhere)
+
+```
+effective ruleset = ~/.claude/fux/global/   (cross-project best practices)
+                  ⊕ ~/.claude/fux/packs/*    (opt-in shareable domain packs)
+                  ⊕ ./.fux/rules/            (this project's domain rules)
+```
+
+`project` overrides `pack` overrides `global`. `fux check` flags conflicts
+instead of silently shadowing.
+
+## Guarantee
+
+Every maintenance command is shell/AST/parse — **no LLM calls**. The only paths
+that call the LLM are the `plan` / `adr` skills, and they ride the session you are
+already in (no background spend). The same "$0, deterministic" promise that made
+graphify trustworthy.
+
+## License
+
+MIT — see [LICENSE](LICENSE).

fux_engine-0.1.0/fux/__init__.py

@@ -0,0 +1,6 @@
+"""Fux — a portable, Claude-aware knowledge engine.
+
+One frontmatter substrate → derived index, graph, and memory views.
+$0 deterministic maintenance; no mandatory LLM calls. See docs/fux-plan.md.
+"""
+__version__ = "0.1.0"

fux_engine-0.1.0/fux/__main__.py

@@ -0,0 +1,4 @@
+from fux.cli import main
+
+if __name__ == "__main__":
+    raise SystemExit(main())

fux_engine-0.1.0/fux/assets/fux-icon.svg

@@ -0,0 +1,61 @@
+<svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 120 120" width="120" height="120">
+  <defs>
+    <!-- tile backgrounds -->
+    <linearGradient id="ic-bg" x1="0" y1="0" x2=".6" y2="1">
+      <stop offset="0%"   stop-color="#16140f"/>
+      <stop offset="100%" stop-color="#0a0908"/>
+    </linearGradient>
+    <radialGradient id="ic-glow" cx="50%" cy="18%" r="70%">
+      <stop offset="0%"   stop-color="#9ed94a1a"/>
+      <stop offset="100%" stop-color="rgba(0,0,0,0)"/>
+    </radialGradient>
+    <linearGradient id="ic-shine" x1="0" y1="0" x2="0" y2="1">
+      <stop offset="0%"   stop-color="rgba(255,255,255,.06)"/>
+      <stop offset="100%" stop-color="rgba(255,255,255,0)"/>
+    </linearGradient>
+    <!-- mark gradient -->
+    <linearGradient id="ic-mg" x1="0" y1="0" x2="1" y2="1">
+      <stop offset="0%"   stop-color="#D2F58F"/>
+      <stop offset="55%"  stop-color="#8FD13F"/>
+      <stop offset="100%" stop-color="#4C8A1B"/>
+    </linearGradient>
+    <clipPath id="ic-mc">
+      <rect x="12" y="10" width="40" height="44" rx="10"/>
+    </clipPath>
+    <filter id="ic-drop" x="-30%" y="-30%" width="160%" height="160%">
+      <feGaussianBlur in="SourceGraphic" stdDeviation="2.5" result="blur"/>
+      <feComposite in="SourceGraphic" in2="blur" operator="over"/>
+    </filter>
+  </defs>
+
+  <!-- tile base -->
+  <rect width="120" height="120" rx="28" fill="url(#ic-bg)"    stroke="rgba(255,255,255,.08)" stroke-width="1"/>
+  <!-- accent glow overlay -->
+  <rect width="120" height="120" rx="28" fill="url(#ic-glow)"/>
+  <!-- top-edge shine -->
+  <rect width="120" height="60"  rx="28" fill="url(#ic-shine)"/>
+
+  <!-- mark: size=78, centred → offset=(120-78)/2=21 -->
+  <svg x="21" y="21" width="78" height="78" viewBox="0 0 64 64">
+    <defs>
+      <linearGradient id="ic-mg2" x1="0" y1="0" x2="1" y2="1">
+        <stop offset="0%"   stop-color="#D2F58F"/>
+        <stop offset="55%"  stop-color="#8FD13F"/>
+        <stop offset="100%" stop-color="#4C8A1B"/>
+      </linearGradient>
+      <clipPath id="ic-mc2">
+        <rect x="12" y="10" width="40" height="44" rx="10"/>
+      </clipPath>
+    </defs>
+    <g filter="url(#ic-drop)">
+      <rect x="12" y="10" width="40" height="44" rx="10" fill="url(#ic-mg2)"/>
+      <g clip-path="url(#ic-mc2)">
+        <polygon points="12,10 32,10 12,30" fill="#fff" opacity=".15"/>
+        <polygon points="52,54 52,34 32,54" fill="#000" opacity=".17"/>
+      </g>
+      <path d="M26 10 L38 10 L38 33 L32 28 L26 33 Z" fill="#000" opacity=".26"/>
+      <rect x="19" y="42" width="26" height="3" rx="1.5" fill="#000" opacity=".24"/>
+      <rect x="19" y="48" width="17" height="3" rx="1.5" fill="#000" opacity=".24"/>
+    </g>
+  </svg>
+</svg>

fux_engine-0.1.0/fux/assets/fux-lockup.svg

@@ -0,0 +1,69 @@
+<svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 440 110" width="440" height="110">
+  <defs>
+    <!-- background -->
+    <linearGradient id="lk-bg" x1="0" y1="0" x2="0" y2="1">
+      <stop offset="0%"   stop-color="#121110"/>
+      <stop offset="100%" stop-color="#0a0a09"/>
+    </linearGradient>
+    <!-- vertical rule -->
+    <linearGradient id="lk-rule" x1="0" y1="0" x2="0" y2="1">
+      <stop offset="0%"   stop-color="rgba(255,255,255,0)"/>
+      <stop offset="50%"  stop-color="rgba(255,255,255,.14)"/>
+      <stop offset="100%" stop-color="rgba(255,255,255,0)"/>
+    </linearGradient>
+    <!-- mark gradient -->
+    <linearGradient id="lk-mg" x1="0" y1="0" x2="1" y2="1">
+      <stop offset="0%"   stop-color="#D2F58F"/>
+      <stop offset="55%"  stop-color="#8FD13F"/>
+      <stop offset="100%" stop-color="#4C8A1B"/>
+    </linearGradient>
+    <clipPath id="lk-mc">
+      <rect x="12" y="10" width="40" height="44" rx="10"/>
+    </clipPath>
+    <filter id="lk-glow" x="-40%" y="-40%" width="180%" height="180%">
+      <feGaussianBlur in="SourceGraphic" stdDeviation="3" result="blur"/>
+      <feComposite in="SourceGraphic" in2="blur" operator="over"/>
+    </filter>
+    <style>
+      @import url('https://fonts.googleapis.com/css2?family=Space+Grotesk:wght@700&amp;family=Space+Mono&amp;display=swap');
+      .lk-eyebrow { font: 8.5px 'Space Mono', 'Courier New', monospace; letter-spacing: .34em; fill: #7a756b; }
+      .lk-name    { font: 700 30px 'Space Grotesk', 'Helvetica Neue', Arial, sans-serif; letter-spacing: .16em; fill: #f2efe9; }
+      .lk-role    { font: 10px 'Space Mono', 'Courier New', monospace; letter-spacing: .26em; fill: #9ed94a; }
+    </style>
+  </defs>
+
+  <!-- plate -->
+  <rect width="440" height="110" rx="14" fill="url(#lk-bg)" stroke="rgba(255,255,255,.07)" stroke-width="1"/>
+
+  <!-- mark: 74×74, vertically centred → y=18 -->
+  <svg x="34" y="18" width="74" height="74" viewBox="0 0 64 64" overflow="visible">
+    <defs>
+      <linearGradient id="lk-mg2" x1="0" y1="0" x2="1" y2="1">
+        <stop offset="0%"   stop-color="#D2F58F"/>
+        <stop offset="55%"  stop-color="#8FD13F"/>
+        <stop offset="100%" stop-color="#4C8A1B"/>
+      </linearGradient>
+      <clipPath id="lk-mc2">
+        <rect x="12" y="10" width="40" height="44" rx="10"/>
+      </clipPath>
+    </defs>
+    <g filter="url(#lk-glow)">
+      <rect x="12" y="10" width="40" height="44" rx="10" fill="url(#lk-mg2)"/>
+      <g clip-path="url(#lk-mc2)">
+        <polygon points="12,10 32,10 12,30" fill="#fff" opacity=".15"/>
+        <polygon points="52,54 52,34 32,54" fill="#000" opacity=".17"/>
+      </g>
+      <path d="M26 10 L38 10 L38 33 L32 28 L26 33 Z" fill="#000" opacity=".26"/>
+      <rect x="19" y="42" width="26" height="3" rx="1.5" fill="#000" opacity=".24"/>
+      <rect x="19" y="48" width="17" height="3" rx="1.5" fill="#000" opacity=".24"/>
+    </g>
+  </svg>
+
+  <!-- vertical rule: x=130, margin 22px top+bottom → y=22..88 -->
+  <rect x="130" y="22" width="1" height="66" fill="url(#lk-rule)"/>
+
+  <!-- text block: x=153, centred on y=55 -->
+  <text x="153" y="30"  class="lk-eyebrow">ALPHA FORGE</text>
+  <text x="153" y="68"  class="lk-name">FUX</text>
+  <text x="153" y="88"  class="lk-role">KNOWLEDGE INDEX</text>
+</svg>

fux_engine-0.1.0/fux/assets/fux-mark.svg

@@ -0,0 +1,26 @@
+<svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 64 64" width="64" height="64">
+  <defs>
+    <linearGradient id="fux-g" x1="0" y1="0" x2="1" y2="1">
+      <stop offset="0%"   stop-color="#D2F58F"/>
+      <stop offset="55%"  stop-color="#8FD13F"/>
+      <stop offset="100%" stop-color="#4C8A1B"/>
+    </linearGradient>
+    <clipPath id="fux-c">
+      <rect x="12" y="10" width="40" height="44" rx="10"/>
+    </clipPath>
+    <filter id="fux-glow" x="-30%" y="-30%" width="160%" height="160%">
+      <feGaussianBlur in="SourceGraphic" stdDeviation="2.5" result="blur"/>
+      <feComposite in="SourceGraphic" in2="blur" operator="over"/>
+    </filter>
+  </defs>
+  <g filter="url(#fux-glow)">
+    <rect x="12" y="10" width="40" height="44" rx="10" fill="url(#fux-g)"/>
+    <g clip-path="url(#fux-c)">
+      <polygon points="12,10 32,10 12,30" fill="#fff" opacity=".15"/>
+      <polygon points="52,54 52,34 32,54" fill="#000" opacity=".17"/>
+    </g>
+    <path d="M26 10 L38 10 L38 33 L32 28 L26 33 Z" fill="#000" opacity=".26"/>
+    <rect x="19" y="42" width="26" height="3" rx="1.5" fill="#000" opacity=".24"/>
+    <rect x="19" y="48" width="17" height="3" rx="1.5" fill="#000" opacity=".24"/>
+  </g>
+</svg>