research_notebook: consolidate to single project-root location only

Per discussion 2026-05-26: there is ONE research_notebook/ per project,
at gigantic_project-COPYME/research_notebook/. Per-subproject
research_notebook/ directories are forbidden — they were inconsistent
in practice (phylonames had only a README, genomesDB had loose audit
TSVs at root, trees_species had reversed-word-order subdirs
user_research/ai_research, secretome had none) and the chat-capture
system always wrote to the project-root location anyway.

Convention updates (gigantic_conventions.md):
- §1: rewritten to mandate single project-root research_notebook/,
  forbid per-subproject ones, and document the optional
  subproject-<name>/ scoping pattern inside research_user/ and
  research_ai/ for per-subproject scratch.
- §25: clarified that "research_user ships empty / research_ai ships
  with docs" applies only to the project-root location.

Disk changes:
- Moved genomesDB's 4 audit files (ncbi_md5_verification, ncbi_status_sweep,
  source_provenance_audit .md + .tsv) from
  subprojects/genomesDB/research_notebook/ → research_notebook/research_user/subproject-genomesDB/
  (sandbox content gitignored — these were never in version control)
- Deleted per-subproject research_notebook/ from phylonames + genomesDB
  + trees_species (trees_species had only zero-byte .gitkeep files in
  reversed-word-order user_research/ and ai_research/ dirs; phylonames
  had only a docs README)

Doc updates across 11 files:
- phylonames + genomesDB + trees_species subproject README and AI_GUIDE
  directory trees no longer list a subproject-level research_notebook/;
  replaced with inline note pointing at project-root sandbox.
- genomesDB STEP_1-sources README + AI_GUIDE: rewrote sections that
  incorrectly implied per-STEP research_notebook/ existence.
- genomesDB STEP_1 workflow README + ai/AI_GUIDE: fixed broken relative
  paths (was ../../research_notebook/... which from workflow level
  pointed at STEP level; corrected to ../../../../research_notebook/...
  reaching project root).
- References to bare research_notebook/research_user/ left as-is where
  unambiguous — §1 mandates the single project-root location, so the
  bare path is unambiguously that.

Note: chat-capture filename format and TRANSCRIPT_CAPTURE_LOG.md
schema unchanged. The PreCompact hook and Save Chat! script were
already writing only to the project-root location.

Author: 000specific

gigantic_project-COPYME/ai/ai_FYIs/gigantic_conventions.md

@@ -8,25 +8,49 @@ At the end of cleanup, this is the source for writing the top-level `GIGANTIC/`
 
 ---
 
-## 1. `research_notebook/` is a sandbox, outside GIGANTIC conventions
+## 1. `research_notebook/` is a sandbox, outside GIGANTIC conventions — and there is ONE, at project root only
 
-`research_notebook/` directories at any level (project, subproject, BLOCK/STEP) are
-personal sandboxes. They have:
+There is **exactly one** `research_notebook/` per project, at the
+project root: `gigantic_project-COPYME/research_notebook/`. Per-subproject
+`research_notebook/` directories are forbidden. (Earlier templates
+allowed per-subproject and per-BLOCK/STEP sandboxes; that pattern was
+consolidated — the per-subproject dirs were inconsistent in practice
+and the chat-capture system always wrote to the project-root location
+anyway.)
+
+The project-root sandbox has:
 
 - **No structure requirements** — any files, any naming, any subdirectory layout
 - **No naming conventions** — users organize however they like
 
-The only hard rule: **GIGANTIC must never pull from anything inside a
-`research_notebook/`.** Workflows, scripts, and pipelines treat these directories as
+The only hard rule: **GIGANTIC must never pull from anything inside
+`research_notebook/`.** Workflows, scripts, and pipelines treat it as
 invisible.
 
-If a user wants GIGANTIC to use a file that lives in their `research_notebook/`,
-they symlink it into `gigantic_project-COPYME/INPUT_user/` (the canonical
-project-level input directory). GIGANTIC reads from `INPUT_user/`, never from
+If a user wants GIGANTIC to use a file that lives in
+`research_notebook/`, they symlink it into
+`gigantic_project-COPYME/INPUT_user/` (the canonical project-level
+input directory). GIGANTIC reads from `INPUT_user/`, never from
 `research_notebook/`.
 
-This separation is what lets `research_notebook/` stay completely free-form while
-GIGANTIC's reproducibility guarantees remain intact.
+This separation is what lets `research_notebook/` stay completely
+free-form while GIGANTIC's reproducibility guarantees remain intact.
+
+**Per-subproject scoping** (when useful): organize inside the project-root
+sandbox with `subproject-<name>/` subdirs, e.g.:
+
+```
+research_notebook/
+├── research_user/
+│   ├── subproject-genomesDB/   # user's per-subproject scratch
+│   ├── subproject-phylonames/
+│   └── species70/              # or any user-defined organization
+└── research_ai/
+    ├── sessions/               # chat captures (per §9)
+    └── subproject-phylonames/  # AI-side per-subproject scratch (when needed)
+```
+
+But this is a convenience pattern, not a requirement — the sandbox is free-form.
 
 ---
 
@@ -510,7 +534,9 @@ All four are in `.gitignore`. When clearing a stuck run, deleting these
 
 ## 25. `research_user/` ships as empty directory; `research_ai/` ships with documentation
 
-Refines §1:
+Refines §1 (and applies only to the single project-root
+`research_notebook/` — per-subproject `research_notebook/` directories
+are forbidden):
 
 - `research_notebook/research_user/` is the user's wild-west sandbox. The
   template ships only the empty directory (`.gitkeep` only — no README,

gigantic_project-COPYME/subprojects/genomesDB/AI_GUIDE.md

@@ -267,7 +267,9 @@ genomesDB/
 │   ├── STEP_3-databases/
 │   └── STEP_4-create_final_species_set/
 │
-├── research_notebook/                       # Personal workspace + AI session captures (§1, §25)
+│   (no per-subproject research_notebook/ — single project-root sandbox at
+│   gigantic_project-COPYME/research_notebook/ per §1, §25; chat captures
+│   land at research_notebook/research_ai/sessions/ per §9)
 │
 ├── STEP_1-sources/
 │   ├── README.md

gigantic_project-COPYME/subprojects/genomesDB/README.md

@@ -219,7 +219,9 @@ genomesDB/
 │   ├── STEP_3-databases/
 │   └── STEP_4-create_final_species_set/    # The final species set canonically lives here
 │
-├── research_notebook/                       # Personal workspace + AI session captures (§1, §25)
+│   (no per-subproject research_notebook/ — single project-root sandbox at
+│   gigantic_project-COPYME/research_notebook/ per §1, §25; chat captures
+│   land at research_notebook/research_ai/sessions/ per §9)
 │
 ├── STEP_1-sources/
 │   ├── README.md

gigantic_project-COPYME/subprojects/genomesDB/STEP_1-sources/AI_GUIDE.md

@@ -47,12 +47,12 @@ Unlike other GIGANTIC subprojects that automatically download or generate data,
 - Users may have custom or unpublished genomes
 - Complete control over what enters GIGANTIC
 
-**research_notebook/research_user/ directory**:
-- Located at `research_notebook/research_user/`
-- User's personal workspace for source data
-- NOT part of GIGANTIC (only README is version-controlled)
-- Can contain anything: raw downloads, scripts, notes, analyses
-- Workflow reads FROM here but doesn't manage it
+**User sandbox for source data** (per §1, §25):
+- Single project-root location: `gigantic_project-COPYME/research_notebook/research_user/`
+- (No per-subproject `research_notebook/` exists.)
+- Wild-west — any files, any organization
+- NOT part of GIGANTIC (contents are gitignored; never version-controlled)
+- Workflow reads FROM here (via INPUT_user symlinks per §17, §18) but doesn't manage it
 
 ---
 
@@ -189,7 +189,7 @@ Workflow run logs are saved to each workflow's `ai/logs/` directory. AI sessions
 | `workflow-*/INPUT_user/source_manifest.tsv` | 4-column manifest | **YES** (required) |
 | `workflow-*/START_HERE-user_config.yaml` | Project name, options | **YES** (project name) |
 | `output_to_input/STEP_1-sources/T1_proteomes/` | Symlinks for STEP_2 | No (auto-created) |
-| `research_notebook/research_user/` | User's personal source data | Personal space |
+| `../../../research_notebook/research_user/` (project-root sandbox per §1, §25) | User's personal source data | Personal space (gitignored) |
 
 ---
 
@@ -265,11 +265,11 @@ STEP_1-sources/
 ├── README.md                    # Human-readable overview
 ├── AI_GUIDE.md          # THIS FILE
 ├── output_to_input/             # Symlinks passed to STEP_2
-│   └── proteomes/               # Created by workflow
+│   └── STEP_1-sources/T1_proteomes/  # Created by workflow
 │       ├── Species1.aa -> ...
 │       └── proteome_manifest.tsv
-├── research_notebook/research_user/               # User's personal workspace
-│   └── README.md                # Only this is part of GIGANTIC
+│   (no per-STEP research_notebook/ — single project-root sandbox at
+│   gigantic_project-COPYME/research_notebook/ per §1, §25)
 └── workflow-COPYME-ingest_source_data/
     ├── README.md
     ├── RUN-workflow.sh

gigantic_project-COPYME/subprojects/genomesDB/STEP_1-sources/README.md

@@ -132,20 +132,24 @@ See `workflow-COPYME-ingest_source_data/README.md` for detailed instructions.
 
 ---
 
-## research_notebook/research_user/ Directory
+## Personal workspace for source data
 
-The `research_notebook/research_user/` directory is the user's personal workspace (per gigantic_conventions §1, §25):
-- Store source genomes/proteomes before ingestion
-- Keep custom scripts, analyses, notes
-- Organize however you want
+Per conventions §1 + §25, the user's wild-west sandbox is the single
+project-root location `gigantic_project-COPYME/research_notebook/research_user/`
+(no per-subproject `research_notebook/` exists). Use it for:
+- Source genomes/proteomes before ingestion
+- Custom scripts, analyses, notes
+- Anything else, organized however you want
 
-**NOT part of GIGANTIC** - only the README is version-controlled.
+The sandbox is gitignored — content is NOT version-controlled.
 
 ---
 
-## Research Notebook
+## Run logs vs AI session logs
 
-Workflow run logs are saved to each workflow's `ai/logs/` directory. AI sessions are extracted project-wide to `research_notebook/research_ai/sessions/`.
+Workflow run logs are saved to each workflow's `ai/logs/` directory.
+AI chat session transcripts (per §9) are captured project-wide at
+`gigantic_project-COPYME/research_notebook/research_ai/sessions/`.
 
 ---
 
@@ -162,9 +166,9 @@ STEP_1-sources/
 ├── README.md                    # This file
 ├── AI_GUIDE.md          # Guide for AI assistants
 ├── output_to_input/             # Symlinks passed to STEP_2
-│   └── proteomes/               # Created by workflow
-├── research_notebook/research_user/               # Your personal workspace (not part of GIGANTIC)
-│   └── README.md                # Only this README is in GIGANTIC
+│   └── STEP_1-sources/T1_proteomes/               # Created by workflow
+│   (no per-STEP research_notebook/ — single project-root sandbox at
+│   gigantic_project-COPYME/research_notebook/ per §1, §25)
 └── workflow-COPYME-ingest_source_data/  # The ingestion workflow template
     ├── README.md
     ├── RUN-workflow.sh

gigantic_project-COPYME/subprojects/genomesDB/STEP_1-sources/workflow-COPYME-ingest_source_data/README.md

@@ -21,7 +21,7 @@ Ingest user-provided genome, proteome, and annotation files into GIGANTIC for do
 
 ## Quick Start
 
-1. Place your source files in the project-level `INPUT_user/genomic_resources/` subdirectories (genomes/, proteomes/, annotations/) or somewhere else accessible (e.g., `../../research_notebook/research_user/`)
+1. Place your source files in the project-level `INPUT_user/genomic_resources/` subdirectories (genomes/, proteomes/, annotations/) or somewhere else accessible (e.g., the project-root sandbox at `../../../../research_notebook/research_user/` per §1, §25)
 
 2. **Ensure files follow GIGANTIC naming convention**:
    ```

gigantic_project-COPYME/subprojects/genomesDB/STEP_1-sources/workflow-COPYME-ingest_source_data/ai/AI_GUIDE.md

@@ -136,7 +136,7 @@ Users must have data files in the project-level `INPUT_user/genomic_resources/`
 - `../../../../INPUT_user/genomic_resources/proteomes/` - Project-level proteome files (.aa)
 - `../../../../INPUT_user/genomic_resources/annotations/` - Project-level annotation files (.gff3/.gtf)
 - `../../../../INPUT_user/genomic_resources/maps/` - Identifier mapping files (.tsv)
-- `../../research_notebook/research_user/` - Personal research data (alternative)
+- `../../../../research_notebook/research_user/` - Project-root sandbox (alternative, per §1, §25)
 - Any accessible path on the system
 
 ### Step 2: Ensure Files Follow GIGANTIC Naming Convention
@@ -367,11 +367,17 @@ bash ai/scripts/003_ai-bash-create_output_symlinks.sh \
 
 ---
 
-## What's NOT in research_notebook/research_user/
+## On the user sandbox (research_notebook/research_user/)
 
-**Important**: The `research_notebook/research_user/` directory at the STEP_1 level is:
-- NOT part of GIGANTIC (only the README is)
-- User's personal space for source data
+Per conventions §1, §25, the user sandbox lives at the **single
+project-root location**: `gigantic_project-COPYME/research_notebook/research_user/`
+(there is no per-STEP or per-subproject `research_notebook/`). Properties:
+
+- NOT part of GIGANTIC (contents are gitignored; never version-controlled)
+- User's personal space for any source data, scripts, notes
+- Wild-west organization — no naming/structure rules
 - Can contain anything the user wants
 
-The workflow reads FROM user_research (or anywhere else) but stores GIGANTIC-managed copies in OUTPUT_pipeline.
+The workflow reads from `INPUT_user/` (which contains symlinks into the
+project-root sandbox per §17, §18), not directly from the sandbox itself.
+GIGANTIC-managed copies of ingested data live in `OUTPUT_pipeline/`.

gigantic_project-COPYME/subprojects/genomesDB/research_notebook/.gitkeep

@@ -71,7 +71,9 @@ phylonames/
 │   └── maps/                                   # Convenience symlink to most recent STEP
 │       └── [project]_map-genus_species_X_phylonames.tsv      (symlink)
 │
-├── research_notebook/                          # Personal workspace + AI session captures (§1, §25)
+│   (no per-subproject research_notebook/ — single project-root sandbox at
+│   gigantic_project-COPYME/research_notebook/ per §1, §25; chat captures
+│   land at research_notebook/research_ai/sessions/ per §9)
 │   ├── research_user/                          # User sandbox (ships empty)
 │   └── research_ai/sessions/                   # Captured chat transcripts (§9)
 │

gigantic_project-COPYME/subprojects/phylonames/AI_GUIDE.md

@@ -204,9 +204,9 @@ phylonames/
 │   └── maps/                                # Convenience landing dir; the .tsv inside is a symlink
 │       └── [project]_map-genus_species_X_phylonames.tsv      (symlink, updated to whichever STEP ran last)
 │
-├── research_notebook/                       # Sandbox + AI session captures (§1, §9, §25)
-│   ├── research_user/                       # User-controlled (ships empty)
-│   └── research_ai/sessions/                # Captured chat transcripts (.jsonl.gz)
+│   (no per-subproject research_notebook/ — the single project-root one
+│   at gigantic_project-COPYME/research_notebook/ serves all subprojects;
+│   see conventions §1, §9, §25)
 │
 ├── STEP_1-generate_and_evaluate/
 │   ├── AI_GUIDE.md                          # STEP-level AI guide
@@ -260,7 +260,9 @@ phylonames/
 **AI lab-notebook logs**: Each workflow run creates a timestamped log in
 its own `ai/logs/` directory. This documents what the workflow did, when,
 with what inputs, and what it produced. Captured AI chat transcripts live
-separately in `research_notebook/research_ai/sessions/` (project-wide).
+in the single project-root sandbox at
+`gigantic_project-COPYME/research_notebook/research_ai/sessions/` (see
+conventions §1, §9, §25).
 
 ---
 
@@ -344,7 +346,8 @@ If you have custom phylonames to apply:
 
 1. **Stage your overrides via the project-level INPUT_user arena**
    (canonical pattern). Put the real file in your sandbox under
-   `research_notebook/research_user/`, then symlink it into the
+   `gigantic_project-COPYME/research_notebook/research_user/` (the
+   single project-root sandbox per §1), then symlink it into the
    project-level `INPUT_user/phylonames/` slot:
    ```bash
    cd ../../INPUT_user/phylonames