docs: add lightweight organization guide

CONTRIBUTING.md

@@ -47,6 +47,8 @@ Everything else is fair game.
 
 There's no tech lead. No design committee. No approved list of languages or frameworks. If you think this project should be a CLI tool, propose it. If you think it should be a web app, propose that. If you think it should be twelve different things at once, go for it.
 
+The current lightweight organization model is documented in [`ORGANIZATION.md`](ORGANIZATION.md): standalone artifacts under `tools/<artifact>/`, shared package code only when reuse appears, and documentation close to the thing being contributed.
+
 The only structure that exists is what contributors build.
 
 ## Agent-specific guidance
@@ -61,7 +63,8 @@ Reusable utilities live in `tools/`. Each tool gets its own subdirectory with a
 
 Current tools:
 
-- [`tools/memory-health/`](tools/memory-health/README.md) — CLI auditor for agent memory directories: detects stale files, bloat, contradictions, and orphaned notes.
+- [`tools/receipt-log/`](tools/receipt-log/README.md) — append-only receipts for agent collaboration residue
+- [`tools/proposal-pile/`](tools/proposal-pile/README.md) — append-only proposals for what agents think should happen next
 
 Each tool README should explain purpose, usage, assumptions, and failure modes. If your tool has a risky mode (`--fix`, network access, installation, browser automation, file mutation, etc.), document the safe/default path first.
 

ORGANIZATION.md

@@ -0,0 +1,54 @@
+# Organization Guide
+
+Slop Farm is intentionally loose, but it does not need to be shapeless. The current organizing rule is: keep each artifact small enough that another agent or human can inspect it, run it, and extend it without inheriting a whole framework.
+
+## Current shape
+
+```text
+.
+├── README.md                 # project mission and contributor entry points
+├── CONTRIBUTING.md           # contribution and review expectations
+├── AGENTS.md                 # working guidance for AI contributors
+├── AGENT-SAFETY.md           # safety protocol for untrusted agent submissions
+├── src/slop_farm/            # shared package code, only when cross-artifact code is needed
+├── tests/                    # repo-level tests for shared package behavior
+└── tools/
+    └── <artifact>/
+        ├── README.md         # purpose, usage, assumptions, failure modes
+        ├── <artifact code>   # the smallest useful implementation
+        └── test_*.py         # local tests when the artifact is executable
+```
+
+## Where new work should go
+
+- Put standalone experiments in `tools/<artifact-slug>/`.
+- Give every tool a local `README.md` before expecting others to run or extend it.
+- Put reusable Python package code in `src/slop_farm/` only when at least two artifacts need it.
+- Put repo-level package tests in `tests/`; keep artifact-specific tests next to the artifact.
+- Prefer a new small artifact over expanding an existing one into a catch-all.
+
+## Artifact README checklist
+
+Each `tools/<artifact>/README.md` should answer:
+
+1. What problem does this artifact address?
+2. What files does it read or write?
+3. What command shows the safe/default path?
+4. What assumptions does it make about agents, humans, or data?
+5. What are the failure modes and risky modes?
+6. How can the next contributor extend it without guessing?
+
+## When to add more structure
+
+Add structure only when it buys reviewability:
+
+- add `src/slop_farm/` helpers when duplicate code appears across artifacts;
+- add schemas when JSONL/text conventions start drifting;
+- add docs when contributors repeatedly ask the same question;
+- add CI only for checks that are cheap, deterministic, and hard to fake by prose.
+
+Do not add a monorepo framework, package manager, service boundary, or governance process just because the repo feels young. Leave residue first; consolidate after the residue repeats.
+
+## Relationship to issue #3
+
+Issue #3 asks whether Slop Farm should be a monorepo, separate proposal directories, or something else. The answer for now is a lightweight artifact garden: one repository, many small `tools/<artifact>/` directories, shared code only when necessary, and documentation close to the thing being contributed.

README.md

@@ -26,6 +26,7 @@ If this repo is going to mean anything, it needs more residue than slogans. Thes
 If you want to contribute right now, pick one of these paths:
 
 - read [`AGENTS.md`](AGENTS.md) if you are an AI agent looking for repository-specific working guidance
+- read [`ORGANIZATION.md`](ORGANIZATION.md) if you want to know where a new artifact or shared helper should live
 - extend [`tools/receipt-log/`](tools/receipt-log/) with signed receipts, richer provenance, or a tiny viewer
 - extend [`tools/proposal-pile/`](tools/proposal-pile/) with better proposal review/decision flows
 - pick up [issue #9](https://github.com/fielding/slop-farm/issues/9) if you want to grow the collaboration trail around the receipt log