Repository navigation
fix(docs): validate aggregate provider counts across published surfaces - #16762
Merged
Merged
Conversation
This was referenced Oct 16, 2026
diegosouzapw
marked this pull request as ready for review
October 16, 2026 03:37
This was referenced Oct 16, 2026
diegosouzapw
added a commit
that referenced
this pull request
Oct 16, 2026
Companion of #16762 (merge-gates §8). That gate landed on `release/v3.8.52`, where the catalog total is **372**. This PR ports the same validator onto `main`, where `docs/reference/PROVIDER_REFERENCE.md` still records **358**. It does not copy 372 onto main. What changed: - `makeProviderCountValidator` in `scripts/check/check-docs-counts-sync.mjs`, required on `README.md`, `AGENTS.md`, and `llm.txt`, and optional on the secondary docs and skill surfaces. - Aggregate claims the checker flagged now say 358. The versioned snapshot “351 registered providers in OmniRoute v3.8.0” and the chat-model registry denominator stay exempt. - Strict `llm.txt` mirrors (`docs/i18n/*/llm.txt`) get the same `352-provider` → `358-provider` token so `check:docs-sync` stays green. - Other locale mirrors were not retranslated. `.i18n-state.json` records the new English source hashes only. Locale hashes stay on the previous source, so `i18n:check` keeps those as non-blocking stale warnings. - Skill surfaces covered by the new check move 327 → 358: `src/shared/constants/agentSkills.ts`, `skills/README.md`, `skills/omni-auth/SKILL.md`, `skills/omni-providers/SKILL.md`. Prettier also realigned existing tables in `docs/architecture/REPOSITORY_MAP.md` and `docs/frameworks/AGENTBRIDGE.md`. - The regression fixture uses `**349** LLM providers` because 358 is the live total on this branch. Checked locally on this branch: provider-count unit test 10/10, `check-docs-counts-sync` exit 0 with no strict drift, `i18n:check` PASS, `check:docs-sync` PASS. Already at `origin/main` (`fd95f823`). CI was still queued at merge time. Local gates on this branch passed.
Contributor
CI Coverage Report
Coverage artifact was not available for this run. |
arashashrafii
pushed a commit
to arashashrafii/OmniRoute
that referenced
this pull request
Oct 24, 2026
…nce#16809) Companion of cheaper-inference#16762 (merge-gates §8). That gate landed on `release/v3.8.52`, where the catalog total is **372**. This PR ports the same validator onto `main`, where `docs/reference/PROVIDER_REFERENCE.md` still records **358**. It does not copy 372 onto main. What changed: - `makeProviderCountValidator` in `scripts/check/check-docs-counts-sync.mjs`, required on `README.md`, `AGENTS.md`, and `llm.txt`, and optional on the secondary docs and skill surfaces. - Aggregate claims the checker flagged now say 358. The versioned snapshot “351 registered providers in OmniRoute v3.8.0” and the chat-model registry denominator stay exempt. - Strict `llm.txt` mirrors (`docs/i18n/*/llm.txt`) get the same `352-provider` → `358-provider` token so `check:docs-sync` stays green. - Other locale mirrors were not retranslated. `.i18n-state.json` records the new English source hashes only. Locale hashes stay on the previous source, so `i18n:check` keeps those as non-blocking stale warnings. - Skill surfaces covered by the new check move 327 → 358: `src/shared/constants/agentSkills.ts`, `skills/README.md`, `skills/omni-auth/SKILL.md`, `skills/omni-providers/SKILL.md`. Prettier also realigned existing tables in `docs/architecture/REPOSITORY_MAP.md` and `docs/frameworks/AGENTBRIDGE.md`. - The regression fixture uses `**349** LLM providers` because 358 is the live total on this branch. Checked locally on this branch: provider-count unit test 10/10, `check-docs-counts-sync` exit 0 with no strict drift, `i18n:check` PASS, `check:docs-sync` PASS. Already at `origin/main` (`fd95f823`). CI was still queued at merge time. Local gates on this branch passed. (cherry picked from commit f295b2e)
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Provider-count claims drifted across the README, secondary documentation, generated skill descriptions and llm.txt mirrors. This updates aggregate claims to the 372 unique provider IDs measured on the validated base and extends the documentation gate to reject mixed current/stale aggregate counts. Explicit historical snapshots, approximate counts and distinct subset denominators remain intact.
Refs #13205.
Validation
Candidate
c36f8d2bb903f93d92fae0bbb24295015ada8142, based on6a1437e6e262db43b50d4bc8cb7b1d658b2e6308:npm run check:docs-allpassed in45.98s. All146 related tests passed in146.81s, with zero failures/skips. The previous runtime-facts timeout is no longer an unresolved candidate check. No deadline or test was relaxed.The current release merge result still needs validation. Draft remains for review/explicit operator approval of protected instruction surfaces (
skills/**/SKILL.md,llm.txtand mirrors). Their changes update factual provider counts only; they do not add commands or operational instructions. No deployment is claimed.