You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
Missing required repository-standard files in MSXOrg/docs #67
Ways-of-Working/Repository-Standard.md defines a required-files table every
repository in the ecosystem must carry. Live-checking MSXOrg/docs against its
own table (via glob, in the repo root and .github/) shows this repository is
missing:
LICENSE
SECURITY.md
SUPPORT.md
CODE_OF_CONDUCT.md
.github/CODEOWNERS
.github/pull_request_template.md
.github/release.yml
(CONTRIBUTING.md was the same kind of gap and is fixed in a companion PR —
this issue is scoped to the remainder.)
Why this is an issue, not a PR
Each of these files requires a real decision rather than boilerplate:
LICENSE — which license applies to this content (docs vs. code license may
differ).
SECURITY.md — a real private-vulnerability-reporting contact/process.
SUPPORT.md — where users should actually ask for help.
CODE_OF_CONDUCT.md — which code of conduct variant to adopt.
.github/CODEOWNERS — who the actual owners/reviewers are.
.github/pull_request_template.md — should match the PR Manager-style title
format already described in Ways-of-Working/PR-Format.md.
.github/release.yml — release-note categories, which should reflect this
repo's own commit/PR conventions.
Per this repo's own main directive, small problems get fixed directly; this
one needs owner decisions first, so it's tracked here instead.
Suggested next step
Decide the content for each file (owner input needed for LICENSE, SECURITY
contact, and CODEOWNERS in particular), then add them in one or a few small
PRs, each following the existing Repository-Standard.md guidance.
Re-checked this list while settling the agent-file set in #140 (#148). It does not change.
The agent files this repository must carry are AGENTS.md, .claude/CLAUDE.md, and .github/copilot-instructions.md, as Agentic Development defines them. All three are present, and no .github/instructions/ file is needed here, so nothing is added to or removed from the seven files above. #148 only moves where the list is written, not what it contains.
One gap the same check surfaced, which this list is missing: .gitattributes is in the required-files table and this repository does not carry it. It is not a decision the way the others are — it normalizes line endings and declares text/binary handling — but the tooling here reads and writes files on both Linux and Windows, and the index and link scripts already have to handle CRLF and LF explicitly, so getting it right matters more than usual. Worth adding to the list, or splitting out as a small standalone change since it needs no owner input.
Correcting my own framing above. I wrote that .gitattributes "is not a decision the way the others are" and "needs no owner input" — that is wrong for this repository, and the "small standalone change" suggestion undersold it.
Three things here need a deliberate answer rather than a default * text=auto:
Binary handling. Two blobs under src/docs/Capabilities/process-psmodule/media/ are images, and they must be declared binary so no normalisation ever touches them.
linguist-* attributes. The site is published from src/, and src/overrides/assets/javascripts/tablesort.min.js is vendored — the Docs workflow already excludes it from linting via FILTER_REGEX_EXCLUDE. Language statistics and diff behaviour should agree with that decision rather than contradict it.
Line endings against the tooling.Update-DocumentationIndex.ps1 and its Pester suite handle CRLF and LF explicitly and are tested both ways, so whatever .gitattributes declares has to stay consistent with what those scripts already do, on Linux and Windows.
None of that is hard, but it is a decision with tests and a published site downstream, which is the same reason the rest of this list is an issue rather than a drive-by commit. Same class as #140, too: the repository that publishes the required-files table not meeting it.
For the record, this gap is specific to MSXOrg/docs. I earlier suggested to a sibling session that it might apply to Storhaug-ting/Kilden as well; it does not — Kilden carries .gitattributes on main.
Activity
MariusStorhaug commentedon Aug 2, 2026
Re-checked this list while settling the agent-file set in #140 (#148). It does not change.
The agent files this repository must carry are
AGENTS.md,.claude/CLAUDE.md, and.github/copilot-instructions.md, as Agentic Development defines them. All three are present, and no.github/instructions/file is needed here, so nothing is added to or removed from the seven files above. #148 only moves where the list is written, not what it contains.One gap the same check surfaced, which this list is missing:
.gitattributesis in the required-files table and this repository does not carry it. It is not a decision the way the others are — it normalizes line endings and declares text/binary handling — but the tooling here reads and writes files on both Linux and Windows, and the index and link scripts already have to handle CRLF and LF explicitly, so getting it right matters more than usual. Worth adding to the list, or splitting out as a small standalone change since it needs no owner input.📖 [Docs]: One page now says which agent files a repository carries (#148
MariusStorhaug commentedon Aug 2, 2026
Correcting my own framing above. I wrote that
.gitattributes"is not a decision the way the others are" and "needs no owner input" — that is wrong for this repository, and the "small standalone change" suggestion undersold it.Three things here need a deliberate answer rather than a default
* text=auto:src/docs/Capabilities/process-psmodule/media/are images, and they must be declared binary so no normalisation ever touches them.linguist-*attributes. The site is published fromsrc/, andsrc/overrides/assets/javascripts/tablesort.min.jsis vendored — the Docs workflow already excludes it from linting viaFILTER_REGEX_EXCLUDE. Language statistics and diff behaviour should agree with that decision rather than contradict it.Update-DocumentationIndex.ps1and its Pester suite handle CRLF and LF explicitly and are tested both ways, so whatever.gitattributesdeclares has to stay consistent with what those scripts already do, on Linux and Windows.None of that is hard, but it is a decision with tests and a published site downstream, which is the same reason the rest of this list is an issue rather than a drive-by commit. Same class as #140, too: the repository that publishes the required-files table not meeting it.
For the record, this gap is specific to MSXOrg/docs. I earlier suggested to a sibling session that it might apply to Storhaug-ting/Kilden as well; it does not — Kilden carries
.gitattributesonmain.📖 [Docs]: Release notes carry complete consumer adoption evidence (#204)