DM3.2: Add D9 micro-task contribution guide and PR template

.github/PULL_REQUEST_TEMPLATE.md

@@ -0,0 +1,52 @@
+## D9 Task — PR Checklist
+
+**Related issue / scoreboard task:** closes #<!-- issue number -->
+
+**Scoreboard entry:** <!-- link to the DM3 scoreboard row, if applicable -->
+
+---
+
+### What this PR does
+
+<!-- One-sentence summary of the contribution (proof, script, note, mapping …) -->
+
+---
+
+### Checklist
+
+#### Reproducibility
+- [ ] All random seeds are fixed and recorded (pass `--seed <n>` or equivalent).
+- [ ] The exact command used to generate output is pasted below (or in the script header).
+- [ ] Output files (`.csv`, `.json`, `.png`) committed to the correct folder (`scripts/out/`, `outputs/`), **or** too large and linked externally with a checksum.
+
+#### Tests / verification
+- [ ] Script runs end-to-end without error on a clean install (`pip install numpy networkx matplotlib`).
+- [ ] Numerical results match the expected values stated in the issue (or deviations are explained).
+- [ ] Lean proof compiles without `sorry` (if applicable).
+
+#### Data
+- [ ] No raw binary blobs > 10 MB committed; large files use a link + SHA-256 hash.
+- [ ] Data sources are cited (FlyWire API URL, paper DOI, etc.).
+
+#### Formatting (analytic notes)
+- [ ] Equations use the agreed LaTeX/Markdown conventions from [CONTRIBUTING.md](https://github.com/TOTOGT/DM3-lab/blob/main/CONTRIBUTING.md).
+- [ ] File is named `docs/<topic>_<initials>_<YYYY-MM>.md` (or `.pdf` with matching `.tex` source).
+
+#### Scoreboard
+- [ ] PR title starts with the task ID, e.g. `[D9-42] …`.
+- [ ] Scoreboard link or issue reference included above.
+
+---
+
+### Reproduction command
+
+```bash
+# paste the exact command here, e.g.:
+# python scripts/collatz_c9_2_sampling.py --N 1000 --M 50 \
+#   --window-type dyadic --start 1 --end 1024 \
+#   --sample-rate 0.1 --seed 42 --output scripts/out/my_run
+```
+
+### Notes for reviewers
+
+<!-- Any context that helps the reviewer validate the result -->

CONTRIBUTING.md

@@ -1,6 +1,6 @@
 # Contributing to AXLE
 
-Thank you for your interest in contributing to the AXLE community! This guide explains how to add files to the repository — including from your **phone or tablet** using the GitHub mobile app.
+Thank you for your interest in contributing to the AXLE community! This guide covers both the GitHub file-upload workflow (including from mobile) and the specific conventions for **D9 micro-tasks** — numerical scripts, analytic notes, and reproducible pull requests.
 
 ---
 
@@ -61,8 +61,82 @@ You can upload files directly to this repository from your phone without needing
 
 ---
 
+## D9 Micro-Task Contributions
+
+D9 tasks are small, self-contained work items (numerical experiments, analytic notes, Lean proofs) tied to the DM3 scoreboard. Anyone can self-assign a task by commenting on the relevant issue.
+
+### Running Numerical Scripts
+
+All scripts live in `/scripts/` and `/simulations/`. Use Python 3.10+.
+
+**1. Install dependencies**
+
+```bash
+pip install numpy networkx matplotlib
+```
+
+**2. Run a script**
+
+```bash
+# Example — C9.2 conditional sampling
+python scripts/collatz_c9_2_sampling.py \
+  --N 1000 --M 50 \
+  --window-type dyadic --start 1 --end 1024 \
+  --sample-rate 0.1 --seed 42 \
+  --output scripts/out/my_run
+```
+
+Output files (`.csv` and `_summary.json`) are written to `scripts/out/`. Create that directory first if it does not exist:
+
+```bash
+mkdir -p scripts/out
+```
+
+**3. Reproduce a prior result**
+
+Always pass `--seed <integer>` to lock the random state. The seed used in any published result must be recorded in the PR description and the corresponding JSON summary file.
+
+**4. Connectome simulations**
+
+```bash
+python simulations/connectome_loader.py   # synthetic fallback, no extra data needed
+python simulations/simple_to_operator.py
+```
+
+Plots are saved to `outputs/` (created automatically).
+
+---
+
+### LaTeX / Formatting Guidelines for Analytic Notes
+
+Place analytic write-ups in `/docs/` as `.md` files (preferred for version control) or `.pdf` (compiled separately).
+
+**Markdown with inline math (`.md`)**
+
+- Use standard GitHub-flavored Markdown.
+- Inline math: `` $f(n) = \lfloor n/2 \rfloor$ ``
+- Display math: fenced with `$$` on its own line.
+- Label every key equation with a bold tag, e.g. `**(Eq. C9.2)**`.
+
+**LaTeX source (`.tex`)**
+
+- Keep one `.tex` file per note; compile with `pdflatex` or `lualatex`.
+- Use `\label{eq:name}` and `\ref{eq:name}` — never bare equation numbers.
+- Preferred packages: `amsmath`, `amssymb`, `hyperref`.
+- Avoid custom macros that are not defined in the same file.
+- Include a `\bibliographystyle{plain}` entry and a `.bib` file when citing external work.
+
+**Naming convention**
+
+```
+docs/<topic>_<author-initials>_<YYYY-MM>.md
+# e.g. docs/collatz_c9_sampling_PG_2026-04.md
+```
+
+---
+
 ## Questions?
 
-Open an [Issue](https://github.com/TOTOGT/AXLE/issues) or reach out:
+Open an [Issue](https://github.com/TOTOGT/DM3-lab/issues) or reach out:
 
 **Contact:** Pablo Grossi | PabloGrossi@hotmail.com | G6 LLC · Newark, NJ · 2026