This repository contains the public repository quality standard, default
community health files, and self-hosted statistics tooling for projects
maintained by trsdn.
Its language is English, and it ships no localized content. See L01 and L03.
The Repository Quality Standard defines the evidence required for maintainable public repositories. Repository-specific requirements may be stricter, but must not silently weaken this baseline.
standard.yml is the same content in machine-readable form,
generated from the document and kept in sync by a required check.
Repository statistics for P09 are documented in
Repository stats, and account-level cards are documented in
Profile statistics.
Cite the pinned form, never the default branch:
https://github.com/trsdn/.github/blob/v1.3.3/docs/repository-quality-standard.md#p09
Criterion identifiers are append-only and are never reused, so a citation keeps its meaning. See decision 0001.
-
Assess the repository against a published version of the standard. To start from a draft rather than a blank record, read what a script can read:
python3 scripts/assess.py --repo owner/name --out drafts/owner-name
It decides only the criteria a fact settles — the topic, the licence, the
permissionsblocks, the action references — and leaves everything that turns on judgementunknown. See drafting a record. -
Record the result in
.github/conformance.yml, following the record format. To start from a record that already names every criterion, copytemplates/conformance.ymlor generate one:python3 scripts/conformance.py --init --repository /path/to/repository
-
Generate the badge from the record and commit it.
-
Add the
trsdn-standardtopic so the repository appears in the inventory. -
Call the reusable check:
jobs: conformance: uses: trsdn/.github/.github/workflows/conformance.yml@main
The topic marks a repository as assessed, not as passing. The outcome lives only in the record.
gh search repos --owner trsdn --topic trsdn-standard --limit 100 \
--json fullName,isArchivedThe difference between that list and the full account list is the outstanding
assessment backlog. scripts/assess.py exists to make that
number movable.
State: Healthy. Every criterion that applies to this repository passes, and the evidence for each is recorded in the self-assessment. The Published Site profile is not among them: this repository defines the standard rather than shipping a product, and its readers are already inside a repository when they read it.
Publishing a green badge over a known gap would devalue every other badge in the estate, so the badge follows the record and never the other way round.
The card is generated by the tooling this repository publishes and committed to
the stats branch on a schedule. It is not fetched from a third-party rendering
service, which is what P09 requires and why Y02 requires it.
GitHub can use the governance and contribution files in this repository for
public trsdn repositories that do not provide their own versions. A file in an
individual repository takes precedence when project-specific guidance is
needed.
Default files improve consistency, but they do not configure branch protection, CI, dependency updates, vulnerability alerts, or secret scanning in other repositories. Those controls must be enabled and verified per repository.
templates/AGENTS.md— starting point for the agent readiness criteriaG01toG08.templates/conformance.yml— starting point for theB11conformance record, naming every criterion in the catalog. Generated, so it cannot fall behind the criteria it lists.templates/repo-stats/— caller workflow and README snippet for theP09repository statistics card.templates/release-notes/— reference release workflow implementing theR07gate, which publishes the changelog entry for the tag as the release notes.
Agent and contributor guidance for this repository is in
AGENTS.md. Decisions behind the standard are recorded in
docs/decisions/.
From a clean checkout, run the repository validation commands:
python3 scripts/standard.py --check
python3 scripts/conformance.py --check
python3 scripts/links.py
python3 -m unittest discover -s tests -v
npx --yes markdownlint-cli2@0.18.1 "**/*.md"
pipx run ruff==0.14.5 check . && pipx run ruff==0.14.5 format --check .The standard and conformance scripts, the link checker, the draft assessor, and
their tests use the Python standard library only. The statistics generator is
separate runtime tooling and installs requests from requirements.txt when
its workflows run.
Markdown linting and Ruff pin their tool versions and are never added as
dependencies, so every script still runs from a bare Python installation.
This repository collects nothing and stores no user data. Routine validation
resolves markdownlint-cli2 from the npm registry and Ruff from PyPI. The
optional statistics generator contacts the GitHub API to render repository and
account SVG cards. The cards it produces are self-contained and embed no remote
fonts or images, so displaying one observes nobody. See Y01 and Y02.
One credential is configured here beyond the token GitHub issues per workflow
run: STATS_TOKEN, an optional fine-grained personal access token the
statistics workflows use where the run token is not enough, with the access
listed in repo stats. If it is exposed, the maintainer
revokes it in GitHub account settings, issues a replacement, and updates the
repository secret; no other system holds a copy, so nothing else has to be
changed, and until it is replaced the workflows fall back to the run token.
GITHUB_TOKEN is issued and revoked per run and needs no such handling. See
B14.
The content and templates in this repository are available under the MIT License.
Nothing here redistributes third-party code. The documents, templates, scripts,
and workflows are written in this repository; the statistics generator's one
runtime dependency and the validation tools are resolved from their registries
by whoever runs them, and are not vendored, bundled, or published onward. There
are therefore no third-party licence obligations to pass on. See B15.