Documentation hub for the Research Project Template
Forking the template? Start here: the code and prose exemplars each ship a 5-minute walkthrough at projects/templates/template_code_project/docs/forking_guide.md (numerical research) and projects/templates/template_prose_project/docs/forking_guide.md (editorial review). The Active Inference exemplar is documented at projects/templates/template_active_inference/README.md. The AutoResearch exemplar is documented at projects/templates/template_autoresearch_project/README.md. The meta-template exemplar is documented at projects/templates/template_template/README.md. The drift checker that gates your fork against the template's contract is scripts/audit/check_template_drift.py (run uv run python scripts/audit/check_template_drift.py from the repo root).
Quick Reference: Documentation Index | How To Use | Architecture | Publication Runbook | FAQ | GitHub / CI | Canonical Facts
The docs/ directory contains project documentation organized by purpose and audience. This is the central hub for all project documentation beyond code comments.
projects/ is a rotating set: directories are promoted, archived, or moved to projects/working/ over time. The only workspace guaranteed to stay in the tree as the control-positive layout for paths and commands is projects/templates/template_code_project/. For the current discovered list, link _generated/active_projects.md—do not treat any other sibling name as permanent.
Machine-generated snippets (including that authoritative list) live under _generated/. Human-written pages should link there instead of copying project rosters.
Documentation is a view over source-owned facts, not a second place to maintain them. Use the narrowest authoritative source for each kind of claim:
| Claim type | Authoritative source | Documentation rule |
|---|---|---|
| Runtime behavior, defaults, and precedence | Implementation, schema, and tests | Verify the documented command or code path; do not infer behavior from an older guide. |
| Repository counts, public roster, coverage, and stage tables | _generated/ and its named producers |
Link the generated artifact; do not copy volatile literals into prose. |
| Manuscript metrics and statistical results | Project analysis outputs → output/data/manuscript_variables.json → hydrated output/manuscript/ |
Inject {{TOKEN}} values at render time; never hand-copy result numbers into manuscript source. |
| Scholarly claims | Primary sources, project bibliography, and claim/evidence registries where present | Match the strength of the prose to the design and evidence; distinguish observation, association, prediction, and causation. |
| Figures, captions, and accessibility text | The same analysis artifact plus figure/artifact registries | Keep plot values, caption statistics, labels, units, uncertainty, provenance, and alt text consistent and source-bound. |
| Publication or release status | Fresh local gates, hosted CI, publication receipts, and the remote commit | Report these separately; a green local check is not proof of publication or remote synchronization. |
The detailed authoring rules live in
rules/documentation_standards.md, manuscript
syntax in guides/manuscript-semantics.md, and
the reusable deep-review prompt in
prompts/comprehensive-assessment/references/comprehensive-research-software-manuscript-review-prompt.md.
graph TD
subgraph EntryPoints["Entry points"]
README[README.md<br/>Project Overview]
DOC_INDEX[documentation-index.md<br/>Full Index]
HOW_TO[core/how-to-use.md<br/>Usage Guide<br/>12 Skill Levels]
end
subgraph CoreDocs["Core documentation"]
ARCH[core/architecture.md<br/>System Design]
WORKFLOW[core/workflow.md<br/>Development Process]
end
subgraph SkillLevels["Skill-based learning"]
L1[guides/getting-started.md<br/>Levels 1-3: Beginner]
L2[guides/figures-and-analysis.md<br/>Levels 4-6: Intermediate]
L3[guides/testing-and-reproducibility.md<br/>Levels 7-9: Advanced]
L4[guides/extending-and-automation.md<br/>Levels 10-12: Expert]
end
subgraph Operational["Operational"]
PIPELINE[RUN_GUIDE.md<br/>Pipeline Orchestration]
BUILD[operational/build<br/>uv, CI/CD]
TROUBLESHOOT[operational/troubleshooting<br/>Fix Issues]
CONFIG[operational/config<br/>Settings and Performance]
LOGGING[operational/logging<br/>Logging System]
end
subgraph Reference["Reference"]
FAQ[reference/faq.md<br/>Common Questions]
CHEATSHEET[reference/quick-start-cheatsheet.md<br/>Command Reference]
API[reference/api-reference.md<br/>Unified API Docs]
RULES[rules/AGENTS.md<br/>Development Standards]
end
README --> DOC_INDEX
README --> HOW_TO
DOC_INDEX --> CoreDocs
DOC_INDEX --> SkillLevels
DOC_INDEX --> Operational
DOC_INDEX --> Reference
HOW_TO --> L1
L1 --> L2
L2 --> L3
L3 --> L4
ARCH --> WORKFLOW
PIPELINE --> BUILD
BUILD --> TROUBLESHOOT
TROUBLESHOOT --> CONFIG
FAQ --> CHEATSHEET
CHEATSHEET --> API
API --> RULES
| Directory | Purpose | Key Contents |
|---|---|---|
core/ |
Essential documentation | how-to-use.md, architecture.md, workflow.md |
guides/ |
Skill-level guides (1-12) | getting-started, figures-and-analysis, testing, extending |
architecture/ |
System design | two-layer-architecture.md, thin-orchestrator, decision-tree |
usage/ |
Content authoring & patterns | examples, markdown guide, style guide, visualization |
operational/ |
Operational workflows | build/, config/, logging/, troubleshooting/ |
maintenance/ |
Long-horizon ops | private-projects-repo, ci-local, regression, archival, bundle |
reference/ |
Reference materials | api-reference, faq, glossary, cheatsheet, workflows |
modules/ |
Infrastructure modules | modules-guide, scientific simulation, pdf-validation, guides/ |
development/ |
Development & contribution | contributing, security, roadmap, testing/ sub-folder |
best-practices/ |
Best practices | version-control, migration, multi-project, backup-recovery |
prompts/ |
AI prompt templates (see prompts/AGENTS.md) | manuscript, registry cross-refs, literature synthesis, code, test, feature, refactor, docs, infra, validation, assessment |
security/ |
Security & provenance | steganography, hashing, secure execution |
rules/ |
Project Rules | AGENTS, README, testing, manuscript, etc. |
audit/ |
Generated point-in-time audit records | filepath audit snapshot and its producer contract |
plans/ |
Active repo-wide handoffs | scoped plans with acceptance lines; completed history stays in git |
images/ |
Documentation-only visual assets | small, public, referenced diagrams or screenshots |
streams/ |
Livestream & talk notes | timestamped session notes tied to releases or papers |
_generated/ |
Generated facts and indexes | counts, public roster, coverage/status evidence, architecture, publication records, skills index |
../README.md- Project overviewcore/how-to-use.md- Usage guideguides/getting-started.md- Getting started (Levels 1-3)reference/faq.md- Common questions
guides/new-project-setup.md- Complete setup checklist with all pitfallsguides/manuscript-semantics.md- Canonical manuscript syntax (citations, cross-references, sections,{{TOKEN}}substitution) shared by all public template exemplarsarchitecture/thin-orchestrator-summary.md- Script pattern
core/architecture.md- System design overviewarchitecture/two-layer-architecture.md- Full architecture guidearchitecture/thin-orchestrator-summary.md- Pattern detailscore/workflow.md- Development processdevelopment/contributing.md- How to contribute
| Need | Document |
|---|---|
| Get started | core/how-to-use.md |
| Create a new project | guides/new-project-setup.md |
| Understand design | architecture/two-layer-architecture.md |
| See examples | usage/examples.md |
| Find answers | reference/faq.md |
| System reference (PAI) | PAI.md |
| Cloud deployment guide | CLOUD_DEPLOY.md |
| Fix an issue | operational/troubleshooting/ |
| Contribute | development/contributing.md |
| Report security issue | development/security.md |
| Understand modules | modules/modules-guide.md |
| Publish a project | guides/publication-runbook.md |
| Best practices | best-practices/best-practices.md |
| Review accessibility | accessibility.md |
| Security policies | security/README.md |
| Validate docs | ../scripts/audit/lint_docs.py |
Run these from the repository root after changing documentation or a producer that feeds it:
uv run python scripts/audit/lint_docs.py --json --repo-root .
uv run python scripts/audit/audit_documentation.py --format markdown
uv run python scripts/audit/check_template_drift.py --strict
uv run python scripts/docgen/counts.py --check
uv run python scripts/docgen/exemplar_roster.py --check
uv run python scripts/docgen/status_evidence.py --check
uv run python scripts/docgen/api_reference.py --check
uv run python -m infrastructure.skills check
uv run python -m infrastructure.skills check-contractslint_docs.py and the generated-file checks are blocking correctness checks.
audit_documentation.py is a broad RedTeam inventory whose advisories require
review and disposition; its raw count is not itself a release verdict.
AGENTS.md— Documentation hub (docs/)../AGENTS.md— Repository system reference (root)documentation-index.md- Full file indexprompts/README.md- AI prompt templates- Agent skills manifest:
uv run python -m infrastructure.skills write(writes.agents/skill_manifest.jsonat repo root when run) ·uv run python -m infrastructure.skills check— see modules/guides/skills-module.md - Active projects under
../projects/may ship a local docs tree (e.g.../projects/templates/template_code_project/docs/); work-in-progress trees under../projects/working/are not discovered until promoted. Authoritative slugs:_generated/active_projects.md