The documents
Every page on this site was written from a source, and the sources are published here rather than summarised away. Raw markdown is the source of truth; the rendered pages are presentation. Where a source is not published, the reason is stated.
The brief pack this site was built from
Nine documents, prepared 22 August 2026 against the Issues-FS repositories, six module repositories and the flagship consumer, with live PyPI queries and measured node, edge and test counts. Published verbatim, at stable paths.
| Document | What it holds |
|---|---|
| 00 · The brief | The three jobs, the verified package table, the thesis, the real numbers, the build order |
| 01 · Concepts index | 14 concept documents with paths, word counts and newcomer-readiness |
| 02 · Issues-FS-lite | The lite specification, extracted. The source for /lite/ |
| 03 · The data model | Nodes, edges, labels, the five regex primitives, the on-disk layout — from code |
| 04 · What ships | 8 packages against live PyPI, 14 CLI commands with real flags, the three agent surfaces, the live graphs |
| 05 · Site architecture | Page-by-page information architecture with sources and publish status |
| 06 · Boundaries and house style | The graphs.sgit.ai boundary, the correction list, licensing, conventions |
| 07 · Gaps and open questions | Documented-not-built, built-not-documented, the seven open questions, six honest tensions |
| 08 · Source manifest | 38 rows, every path verified, machine-readable |
| README · Licence | The pack's own index and licence scope note |
These are fetchable at stable constructed paths: /briefs/<filename>.md. That is a promise rather than an accident — agents rely on constructed paths, so the convention is stated rather than left to be inferred.
How the sources were tiered
The manifest tiers all 38 verified sources by what can be done with them. Publishing the tiering as well as the result is the point: a reader can see what was excluded and disagree.
| Tier | Count | Means |
|---|---|---|
| Tier 0 | 18 | Publishable near-as-is; highest value |
| Tier 1 | 14 | Publishable with a correction, a framing note, or a supersession note |
| Tier 2 | 4 | Cite, do not republish — third-party imports and curiosities |
| Tier 3 | 2 | Do not publish — superseded, or staging code |
The upstream sources, by section
The corpus itself lives in Issues-FS__Docs (59 documents) and the module repositories. Paths are given so that a reader can go to the source; the documents are not mirrored here, for the licence reason in the next section.
| Used by | Source |
|---|---|
| /lite/ | SGraph-AI__App__Send → team/humans/dinis_cruz/briefs/05/06/email-fs-lite-v0.6.md §7 (lines 625–747), 6 May 2026 · and email-fs-comparison.md (2,927 w) for the “usage profile, not a fork” framing · status of record: team/roles/librarian/reality/ai-agents/index.md:99 |
| /model/ | issues_fs/schemas/graph/Schema__Node.py · Schema__Node__Link.py · Safe_Str__Graph_Types.py · schemas/enums/Enum__Graph__Storage__Backend.py · docs/issues_fs/architecture/0.41.0__issues-fs__memory-fs-abstraction-layer.md |
| /cli/ | Issues-FS__CLI → issues_fs_cli/cli/ (1,080 LOC) and CLI__Output.py |
| /dsl/ | issues_fs/issues/issues_file/ (11 files) · Issues-FS/.issues/*.issues (3 files) · the parser tests |
| /concepts/ | docs/to_classify/v0_4_0__issues-fs__thinking-in-graphs.md (5,013 w, 5 Feb 2026) · …compatibility-through-connectivity.md (3,847 w) · …llm-as-execution-engine.md (3,884 w) · …lexicon-architecture-v2.md (4,485 w) · …use-case-pattern.md (2,909 w) |
| /roles/ | docs/development/guide__agentic-role-based-workflow.md (5,814 w, Status Active) · …role-ecosystem-guide.md (4,475 w) · the four individual role documents |
| /origins/ | docs/to_classify/6-feb-other/v0_4_0__issues-fs__the-journey.md (3,483 w, Status: Historical Record) |
| /examples/ | SGraph-AI__App__Send/.issues/ (84 files) · .issues/config/node-types.json · link-types.json · the three other live graphs |
What is deliberately not published
| What | Words | Why not |
|---|---|---|
| The imported briefing about a graph-database library | 5,375 | Third-party, unattributed. It is about that library rather than about any Issues-FS use of it — because there is no such use. Cited, not republished |
| Seven type-safety briefs about the utilities library Issues-FS builds on | ~24,000 | Third-party, unattributed. With the above, roughly 18% of the corpus carries no source attribution and no licence line |
| Lexicon architecture v1 | 2,509 | Superseded and unmarked. It sits in a folder called already-legacy/ and carries no supersession marker in the file itself — so an agent fetching the path directly has nothing telling it so. Marking it at source is N6 |
The to_refactor-in/ staging directory | 12 files | Un-migrated code, unpackaged and unimportable, still importing the project Issues-FS was extracted from |
| The two screenplays | 6,605 | Catalogued at /curiosities/, held back pending the licence question |
The licence position for anything reproduced here
This site's own content is CC BY 4.0. The material it draws on is not uniformly so, and the site does not paper over the difference:
- The Issues-FS repositories are Apache 2.0, byte-identical across all nine, and none names a copyright holder — every
LICENSEstill carries the unfilledCopyright [yyyy] [name of copyright owner], and no repository has aNOTICE. - The CC BY claim has no backing. One README states that documentation is CC BY 4.0. That sentence is the only occurrence of “CC BY” in the entire corpus, and the repository holding all 59 documents ships Apache 2.0 and says nothing about it.
- Lite crosses the boundary. Its specification is in a CC BY 4.0 repository while the Issues-FS repositories are Apache 2.0. The provenance is stated on the page.
So: paths are given, quotations are attributed, and whole documents are not mirrored until the position is settled. The full note · requests N1 and N2.