loopbench 0.1.0__tar.gz

loopbench-0.1.0/.github/workflows/publish.yml

@@ -0,0 +1,33 @@
+name: Publish to PyPI
+
+on:
+  release:
+    types: [published]
+  workflow_dispatch:
+
+permissions:
+  id-token: write
+
+jobs:
+  publish:
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@v4
+
+      - uses: actions/setup-python@v5
+        with:
+          python-version: "3.12"
+
+      - name: Install build tools
+        run: pip install hatchling build
+
+      - name: Build package
+        run: python -m build
+
+      # Uses PYPI_API_TOKEN when set; otherwise OIDC trusted publishing.
+      - name: Publish to PyPI
+        uses: pypa/gh-action-pypi-publish@release/v1
+        with:
+          password: ${{ secrets.PYPI_API_TOKEN }}
+          skip-existing: true
+

loopbench-0.1.0/.github/workflows/test.yml

@@ -0,0 +1,51 @@
+name: test
+
+on:
+  push:
+    branches: [main, master]
+  pull_request:
+    branches: [main, master]
+
+jobs:
+  test:
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@v4
+
+      - uses: actions/checkout@v4
+        with:
+          repository: KanakMalpani/LoopGym
+          path: deps/LoopGym
+
+      - uses: actions/checkout@v4
+        with:
+          repository: KanakMalpani/Loop-Core-Engineering
+          path: deps/loop-core
+
+      - uses: actions/setup-python@v5
+        with:
+          python-version: "3.12"
+
+      - name: Install loopgym
+        run: pip install -e deps/LoopGym
+
+      - name: Install loopbench
+        run: pip install -e ".[dev]"
+
+      - name: Validate submission LSS specs (Loop Core Engineering)
+        run: |
+          pip install -r deps/loop-core/requirements.txt
+          python deps/loop-core/tools/validate_lss.py submissions/examples/spec-fast-loop.yaml
+          python deps/loop-core/tools/validate_lss.py submissions/examples/spec-thorough-loop.yaml
+
+      - name: Run tests
+        run: pytest tests/ -q
+
+      - name: Validate leaderboard
+        run: loopbench validate leaderboard/entries.json
+
+      - name: Golden submission CLI
+        run: |
+          loopbench run --task LB-CR-1 --spec submissions/examples/spec-fast-loop.yaml \
+            --seeds 0,1 --submitter ci-fast -o /tmp/fast.json
+          loopbench validate /tmp/fast.json

loopbench-0.1.0/.gitignore

@@ -0,0 +1,17 @@
+__pycache__/
+*.py[cod]
+.venv/
+venv/
+.pytest_cache/
+.ruff_cache/
+*.egg-info/
+dist/
+build/
+.DS_Store
+.env
+.env.*
+!.env.example
+*.pem
+credentials.json
+/tmp/
+results*.json

loopbench-0.1.0/CONTRIBUTING.md

@@ -0,0 +1,33 @@
+# Contributing to LoopBench
+
+## What belongs here
+
+- Benchmark task definitions (ALS v2 YAML)
+- Observed LES computation and submission schema
+- CLI, conformance validation, leaderboard entries
+- Golden submission tests
+
+## What does not belong here
+
+- LSS schema — [Loop Core Engineering](https://github.com/KanakMalpani/Loop-Core-Engineering)
+- Environment runtime — [LoopGym](https://github.com/KanakMalpani/LoopGym)
+- Dataset records — [LoopNet](https://github.com/KanakMalpani/loopnet)
+
+## Before opening a PR
+
+```bash
+pip install -e ../LoopGym   # or deps/LoopGym in CI layout
+pip install -e ".[dev]"
+pytest tests/ -q
+loopbench validate leaderboard/entries.json
+```
+
+## Submitting benchmark results
+
+1. Run all tasks: `loopbench run --task LB-CR-1,LB-RS-1,LB-MA-1 --spec your-loop.yaml -o results.json`
+2. Validate: `loopbench validate results.json`
+3. Open a PR adding your entry to `leaderboard/entries.json`
+
+## License
+
+MIT — see [LICENSE](LICENSE).

loopbench-0.1.0/LICENSE

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

loopbench-0.1.0/PKG-INFO

@@ -0,0 +1,182 @@
+Metadata-Version: 2.4
+Name: loopbench
+Version: 0.1.0
+Summary: LoopBench — benchmark suite, metrics, submission pipeline, leaderboards
+Project-URL: Homepage, https://github.com/KanakMalpani/LoopBench
+Project-URL: Repository, https://github.com/KanakMalpani/LoopBench
+Project-URL: Issues, https://github.com/KanakMalpani/LoopBench/issues
+Project-URL: Documentation, https://github.com/KanakMalpani/LoopBench/blob/main/SUITE-OVERVIEW.md
+Project-URL: LoopGym, https://github.com/KanakMalpani/LoopGym
+Project-URL: Loop Core Engineering, https://github.com/KanakMalpani/Loop-Core-Engineering
+Author: Kanak Malpani
+License: MIT
+License-File: LICENSE
+Keywords: benchmark,les,loop-engineering,loopbench
+Classifier: Development Status :: 4 - Beta
+Classifier: License :: OSI Approved :: MIT License
+Classifier: Programming Language :: Python :: 3.12
+Requires-Python: >=3.12
+Requires-Dist: jsonschema>=4.21
+Requires-Dist: pyyaml>=6.0
+Provides-Extra: dev
+Requires-Dist: pytest>=8.0; extra == 'dev'
+Requires-Dist: ruff>=0.4; extra == 'dev'
+Description-Content-Type: text/markdown
+
+<p align="center">
+  <strong>LoopBench</strong><br>
+  <em>MLPerf for loops.</em>
+</p>
+
+<p align="center">
+  <a href="https://github.com/KanakMalpani/LoopBench/actions/workflows/test.yml"><img src="https://github.com/KanakMalpani/LoopBench/actions/workflows/test.yml/badge.svg" alt="CI"></a>
+  <a href="LICENSE"><img src="https://img.shields.io/badge/License-MIT-yellow.svg" alt="MIT"></a>
+  <img src="https://img.shields.io/badge/python-3.12+-blue.svg" alt="Python 3.12+">
+  <a href="SUITE-OVERVIEW.md"><img src="https://img.shields.io/badge/suite-ALS_v2-blue.svg" alt="ALS v2"></a>
+  <img src="https://img.shields.io/badge/tasks-3-green.svg" alt="3 tasks">
+</p>
+
+---
+
+**LoopBench** is the public scoreboard for Loop Engineering — fixed tasks, fixed seeds, observed [LES](https://github.com/KanakMalpani/Loop-Core-Engineering/blob/main/specs/les-1.0.md), and a submission pipeline anyone can audit.
+
+You bring an [LSS](https://github.com/KanakMalpani/Loop-Core-Engineering) loop spec. LoopBench runs it through [LoopGym](https://github.com/KanakMalpani/LoopGym), computes **LES_obs** across eight categories, validates your results JSON, and ranks you on the leaderboard. No hand-waved demos.
+
+```bash
+loopbench run --task LB-CR-1 --spec your-loop.yaml --seeds 0,1,2,3,4 -o results.json
+loopbench validate results.json
+```
+
+<p align="center">
+  <a href="#-run-your-first-score"><strong>Run your first score →</strong></a> ·
+  <a href="leaderboard/entries.json">Leaderboard</a> ·
+  <a href="SUITE-OVERVIEW.md">Suite architecture</a>
+</p>
+
+<p align="center">
+  <img src="assets/demo.gif" alt="LoopBench CLI demo: install, list tasks, run, validate, rank" width="720">
+</p>
+
+---
+
+## The contract
+
+```mermaid
+flowchart LR
+  YOU[Your LSS spec]
+  LB[LoopBench<br/>tasks · scoring · schema]
+  LG[LoopGym<br/>execution]
+  OUT[results.json → leaderboard]
+
+  YOU --> LB
+  LB -->|env_id, seeds| LG
+  LG -->|trajectories| LB
+  LB --> OUT
+```
+
+| Layer | Owns | Repo |
+|-------|------|------|
+| **Spec** | LSS schema, LES formulas | [Loop Core Engineering](https://github.com/KanakMalpani/Loop-Core-Engineering) |
+| **Data** | Trajectories (optional holdout) | [LoopNet](https://github.com/KanakMalpani/loopnet) |
+| **Runtime** | `env.run_episode()` | [LoopGym](https://github.com/KanakMalpani/LoopGym) |
+| **Measurement** | Tasks, LES_obs, submissions | **LoopBench** |
+
+LoopBench **defines** and **scores**. LoopGym **runs**. Never the other way around.
+
+---
+
+## ⚡ Run your first score
+
+```bash
+pip install git+https://github.com/KanakMalpani/LoopGym.git
+pip install git+https://github.com/KanakMalpani/LoopBench.git
+
+loopbench list
+
+loopbench run \
+  --task LB-CR-1 \
+  --spec submissions/examples/spec-fast-loop.yaml \
+  --seeds 0,1,2,3,4 \
+  -o results.json
+
+loopbench validate results.json
+loopbench rank leaderboard/entries.json
+```
+
+**Local dev** (sibling clones):
+
+```bash
+git clone https://github.com/KanakMalpani/LoopGym.git
+git clone https://github.com/KanakMalpani/LoopBench.git
+cd LoopBench && pip install -e ../LoopGym -e ".[dev]"
+```
+
+On Windows: `py -3.12` if needed. PyPI: [PUBLISHING.md](PUBLISHING.md).
+
+---
+
+## Tasks (v0.1 · ALS v2)
+
+| ID | Name | Env | What it stress-tests |
+|----|------|-----|----------------------|
+| `LB-CR-1` | Code repair | `loopbench/code-repair-v1` | Effectiveness, speed, robustness |
+| `LB-RS-1` | Research synthesis | `loopbench/research-synthesis-v1` | Effectiveness, cost |
+| `LB-MA-1` | Multi-agent debate | `loopbench/multi-agent-debate-v1` | Autonomy, scalability |
+
+Each task ships YAML + README under [`tasks/`](tasks/). Five seeds by default. Success@k + **LES_obs** composite.
+
+---
+
+## Metrics
+
+| Metric | Meaning |
+|--------|---------|
+| **Success@k** | Fraction of instances reaching goal threshold `g_target` |
+| **LES_obs** | Observed eight-category composite ∈ `[0, 1]` — see [`metrics/les-compute.md`](metrics/les-compute.md) |
+| **Cost** | Estimated USD per run from LSS cost limits |
+| **Robustness** | Quality retention across seeds |
+
+Display scale `0–100` is optional (`les_display = les_observed × 100`).
+
+---
+
+## Submit to the leaderboard
+
+1. Run all tasks (or start with one):  
+   `loopbench run --task LB-CR-1,LB-RS-1,LB-MA-1 --spec your-loop.yaml -o results.json`
+2. Validate: `loopbench validate results.json`
+3. Open a PR adding your entry to [`leaderboard/entries.json`](leaderboard/entries.json)
+
+v0.1 rankings accept **SimEnv** submissions only (no API keys, fully reproducible). LiveEnv tier: v0.2.
+
+---
+
+## Repository layout
+
+| Path | Purpose |
+|------|---------|
+| [`tasks/`](tasks/) | ALS v2 task definitions |
+| [`metrics/les-compute.md`](metrics/les-compute.md) | LES_obs formulas |
+| [`submit/schema.json`](submit/schema.json) | Submission JSON schema |
+| [`loopbench/`](loopbench/) | Runner, LES compute, conformance |
+| [`leaderboard/`](leaderboard/) | Public rankings (JSON v0.1) |
+| [`submissions/examples/`](submissions/examples/) | Reference specs |
+
+---
+
+## Citation
+
+```bibtex
+@software{loopbench2026,
+  title={LoopBench: Benchmark Suite for Loop Engineering},
+  author={Malpani, Kanak},
+  year={2026},
+  url={https://github.com/KanakMalpani/LoopBench}
+}
+```
+
+---
+
+<p align="center">
+  <sub>MIT · v0.1 · <a href="CONTRIBUTING.md">Contributing</a> · <a href="SECURITY.md">Security</a> · <a href="STATUS.md">Status</a></sub>
+</p>

loopbench-0.1.0/PLAN.md

@@ -0,0 +1,56 @@
+# 06 — loopbench
+
+## One-line purpose
+
+**LoopBench** — world's definitive loop benchmark suite, metrics, submission pipeline, and leaderboards.
+
+## Why this repo exists
+
+Fields need **comparable scores**. LoopBench is MLPerf for loops — tasks, rules, anti-gaming, public ranks.
+
+## Scope (in scope)
+
+- Task definitions (ALS v2): code-repair, research-synthesis, multi-agent-debate (+ expand)
+- Oracle specifications per task
+- Metrics: Success@k, LES-obs, cost, robustness (5 seeds), safety
+- Submission format: container + LSS + results JSON
+- Local runner CLI (`loopbench run`)
+- Leaderboard schema (static JSON v0.1 → web v0.2)
+- Conformance tests: "does submission follow rules?"
+
+## Scope (out of scope)
+
+- Gym implementation → `05-loopgym` (LoopBench *defines*, LoopGym *runs*)
+
+## Deliverables v0.1
+
+- [x] `tasks/` — 3 task specs (YAML + README each)
+- [x] `metrics/les-compute.md` — how observed LES is computed on bench
+- [x] `submit/schema.json` — results file format
+- [x] `cli/loopbench.py` — local eval
+- [x] `leaderboard/` — README + example entries JSON
+
+## Task IDs (initial)
+
+| ID | Name |
+|----|------|
+| `LB-CR-1` | Code repair |
+| `LB-RS-1` | Research synthesis |
+| `LB-MA-1` | Multi-agent debate |
+
+## Dependencies
+
+- **01-loop-engineering-core** — LES, LSS
+- **05-loopgym** — execution engine
+
+## Success criteria
+
+Two different LSS specs submitted locally; leaderboard JSON ranks by composite; CI runs golden submission.
+
+## Agent instructions
+
+Align with existing `benchmarks/` in `Loop Engineering` repo then **move** canonical tasks here. Fixed seeds, pinned deps.
+
+## Status
+
+✅ v0.1 shipped (2026-06-13) — see [STATUS.md](./STATUS.md)

loopbench-0.1.0/PUBLISHING.md

@@ -0,0 +1,42 @@
+# Publishing LoopBench to PyPI
+
+**Prerequisite:** [LoopGym](https://pypi.org/project/loopgym/) must be published first.
+
+## One-time setup
+
+1. Create a PyPI project named **`loopbench`**.
+2. **Preferred:** configure [trusted publishing](https://docs.pypi.org/trusted-publishers/) on the PyPI project:
+   - **PyPI project name:** `loopbench`
+   - **Owner:** `KanakMalpani`
+   - **Repository name:** `LoopBench`
+   - **Workflow name:** `publish.yml`
+   - **Environment name:** *(leave blank)*
+
+3. **Fallback:** add **`PYPI_API_TOKEN`** (upload scope) to this repo's Actions secrets.
+
+## Publish
+
+```bash
+git tag v0.1.0
+git push origin v0.1.0
+gh release create v0.1.0 --title "v0.1.0" --notes "Initial public release"
+```
+
+Or trigger **Actions → Publish to PyPI** manually.
+
+## Install
+
+```bash
+pip install loopbench loopgym
+loopbench list
+```
+
+## Verify locally
+
+```bash
+pip install -e ../LoopGym
+pip install build
+python -m build
+pip install dist/loopbench-*.whl
+pytest tests/ -q
+```

loopbench-0.1.0/README.md

@@ -0,0 +1,157 @@
+<p align="center">
+  <strong>LoopBench</strong><br>
+  <em>MLPerf for loops.</em>
+</p>
+
+<p align="center">
+  <a href="https://github.com/KanakMalpani/LoopBench/actions/workflows/test.yml"><img src="https://github.com/KanakMalpani/LoopBench/actions/workflows/test.yml/badge.svg" alt="CI"></a>
+  <a href="LICENSE"><img src="https://img.shields.io/badge/License-MIT-yellow.svg" alt="MIT"></a>
+  <img src="https://img.shields.io/badge/python-3.12+-blue.svg" alt="Python 3.12+">
+  <a href="SUITE-OVERVIEW.md"><img src="https://img.shields.io/badge/suite-ALS_v2-blue.svg" alt="ALS v2"></a>
+  <img src="https://img.shields.io/badge/tasks-3-green.svg" alt="3 tasks">
+</p>
+
+---
+
+**LoopBench** is the public scoreboard for Loop Engineering — fixed tasks, fixed seeds, observed [LES](https://github.com/KanakMalpani/Loop-Core-Engineering/blob/main/specs/les-1.0.md), and a submission pipeline anyone can audit.
+
+You bring an [LSS](https://github.com/KanakMalpani/Loop-Core-Engineering) loop spec. LoopBench runs it through [LoopGym](https://github.com/KanakMalpani/LoopGym), computes **LES_obs** across eight categories, validates your results JSON, and ranks you on the leaderboard. No hand-waved demos.
+
+```bash
+loopbench run --task LB-CR-1 --spec your-loop.yaml --seeds 0,1,2,3,4 -o results.json
+loopbench validate results.json
+```
+
+<p align="center">
+  <a href="#-run-your-first-score"><strong>Run your first score →</strong></a> ·
+  <a href="leaderboard/entries.json">Leaderboard</a> ·
+  <a href="SUITE-OVERVIEW.md">Suite architecture</a>
+</p>
+
+<p align="center">
+  <img src="assets/demo.gif" alt="LoopBench CLI demo: install, list tasks, run, validate, rank" width="720">
+</p>
+
+---
+
+## The contract
+
+```mermaid
+flowchart LR
+  YOU[Your LSS spec]
+  LB[LoopBench<br/>tasks · scoring · schema]
+  LG[LoopGym<br/>execution]
+  OUT[results.json → leaderboard]
+
+  YOU --> LB
+  LB -->|env_id, seeds| LG
+  LG -->|trajectories| LB
+  LB --> OUT
+```
+
+| Layer | Owns | Repo |
+|-------|------|------|
+| **Spec** | LSS schema, LES formulas | [Loop Core Engineering](https://github.com/KanakMalpani/Loop-Core-Engineering) |
+| **Data** | Trajectories (optional holdout) | [LoopNet](https://github.com/KanakMalpani/loopnet) |
+| **Runtime** | `env.run_episode()` | [LoopGym](https://github.com/KanakMalpani/LoopGym) |
+| **Measurement** | Tasks, LES_obs, submissions | **LoopBench** |
+
+LoopBench **defines** and **scores**. LoopGym **runs**. Never the other way around.
+
+---
+
+## ⚡ Run your first score
+
+```bash
+pip install git+https://github.com/KanakMalpani/LoopGym.git
+pip install git+https://github.com/KanakMalpani/LoopBench.git
+
+loopbench list
+
+loopbench run \
+  --task LB-CR-1 \
+  --spec submissions/examples/spec-fast-loop.yaml \
+  --seeds 0,1,2,3,4 \
+  -o results.json
+
+loopbench validate results.json
+loopbench rank leaderboard/entries.json
+```
+
+**Local dev** (sibling clones):
+
+```bash
+git clone https://github.com/KanakMalpani/LoopGym.git
+git clone https://github.com/KanakMalpani/LoopBench.git
+cd LoopBench && pip install -e ../LoopGym -e ".[dev]"
+```
+
+On Windows: `py -3.12` if needed. PyPI: [PUBLISHING.md](PUBLISHING.md).
+
+---
+
+## Tasks (v0.1 · ALS v2)
+
+| ID | Name | Env | What it stress-tests |
+|----|------|-----|----------------------|
+| `LB-CR-1` | Code repair | `loopbench/code-repair-v1` | Effectiveness, speed, robustness |
+| `LB-RS-1` | Research synthesis | `loopbench/research-synthesis-v1` | Effectiveness, cost |
+| `LB-MA-1` | Multi-agent debate | `loopbench/multi-agent-debate-v1` | Autonomy, scalability |
+
+Each task ships YAML + README under [`tasks/`](tasks/). Five seeds by default. Success@k + **LES_obs** composite.
+
+---
+
+## Metrics
+
+| Metric | Meaning |
+|--------|---------|
+| **Success@k** | Fraction of instances reaching goal threshold `g_target` |
+| **LES_obs** | Observed eight-category composite ∈ `[0, 1]` — see [`metrics/les-compute.md`](metrics/les-compute.md) |
+| **Cost** | Estimated USD per run from LSS cost limits |
+| **Robustness** | Quality retention across seeds |
+
+Display scale `0–100` is optional (`les_display = les_observed × 100`).
+
+---
+
+## Submit to the leaderboard
+
+1. Run all tasks (or start with one):  
+   `loopbench run --task LB-CR-1,LB-RS-1,LB-MA-1 --spec your-loop.yaml -o results.json`
+2. Validate: `loopbench validate results.json`
+3. Open a PR adding your entry to [`leaderboard/entries.json`](leaderboard/entries.json)
+
+v0.1 rankings accept **SimEnv** submissions only (no API keys, fully reproducible). LiveEnv tier: v0.2.
+
+---
+
+## Repository layout
+
+| Path | Purpose |
+|------|---------|
+| [`tasks/`](tasks/) | ALS v2 task definitions |
+| [`metrics/les-compute.md`](metrics/les-compute.md) | LES_obs formulas |
+| [`submit/schema.json`](submit/schema.json) | Submission JSON schema |
+| [`loopbench/`](loopbench/) | Runner, LES compute, conformance |
+| [`leaderboard/`](leaderboard/) | Public rankings (JSON v0.1) |
+| [`submissions/examples/`](submissions/examples/) | Reference specs |
+
+---
+
+## Citation
+
+```bibtex
+@software{loopbench2026,
+  title={LoopBench: Benchmark Suite for Loop Engineering},
+  author={Malpani, Kanak},
+  year={2026},
+  url={https://github.com/KanakMalpani/LoopBench}
+}
+```
+
+---
+
+<p align="center">
+  <sub>MIT · v0.1 · <a href="CONTRIBUTING.md">Contributing</a> · <a href="SECURITY.md">Security</a> · <a href="STATUS.md">Status</a></sub>
+</p>

loopbench-0.1.0/SECURITY.md

@@ -0,0 +1,15 @@
+# Security Policy
+
+## Supported versions
+
+| Version | Supported |
+|---------|-----------|
+| 0.1.x   | Yes       |
+
+## Reporting
+
+Report privately via [GitHub Security Advisories](https://github.com/KanakMalpani/LoopBench/security/advisories/new).
+
+## Submissions
+
+Validate all results with `loopbench validate` before opening leaderboard PRs. v0.1 accepts SimEnv only for public rankings.

loopbench-0.1.0/STATUS.md

@@ -0,0 +1,31 @@
+# Status
+
+| Field | Value |
+|-------|-------|
+| **Phase** | v0.1 shipped |
+| **Symbol** | ✅ |
+| **Started** | 2026-06-13 |
+| **Shipped** | 2026-06-13 |
+| **Owner** | — |
+| **Blockers** | — |
+| **Notes** | Published at https://github.com/KanakMalpani/LoopBench; LiveEnv submissions v0.2 |
+
+## Completion checklist
+
+- [x] `tasks/` — 3 task specs (YAML + README each)
+- [x] `metrics/les-compute.md` — observed LES computation
+- [x] `submit/schema.json` — results file format
+- [x] `cli/loopbench.py` — local eval CLI
+- [x] `leaderboard/` — README + example entries JSON
+- [x] Golden submission tests + CI
+- [x] `SYNC.md`, `LICENSE`, `SUITE-OVERVIEW.md`
+- [x] LSS validation in CI (01-core)
+- [x] Smoke tests for all 3 tasks
+- [ ] PyPI publish (`pip install loopbench` — pending trusted publisher or token on PyPI)
+
+## Links
+
+- Parent workspace: [../README.md](../README.md)
+- LSS/LES: [../01-loop-engineering-core](../01-loop-engineering-core/)
+- Runtime: [../05-loopgym](../05-loopgym/)
+- Agent brief: [../AGENT-BRIEF.md](../AGENT-BRIEF.md)