docs: refresh roadmap for September planning - #6897
Conversation
Publish a theme-led Now/Next/Later snapshot from the September session and keep the August edition in docs/archived-roadmaps/ so monthly archives do not accumulate in one file. Assisted-by: Cursor Grok 4.6 Signed-off-by: Adam Scerra <ascerra@redhat.com> Co-authored-by: Cursor <cursoragent@cursor.com>
PR Summary by QodoRefresh September roadmap and add monthly archives
AI Description
Diagram
High-Level Assessment
Files changed (7)
|
Code Review by Qodo
1.
|
![qodo-code-review[bot]](https://avatars.githubusercontent.com/in/484649?s=60&v=4){kind=small}
|
🤖 Review · Commit: |
Site previewPreview: https://997cc33b-site.fullsend-ai.workers.dev Commit: |
Codecov Report✅ All modified and coverable lines are covered by tests. 📢 Thoughts on this report? Let us know! |
Gate auto-merge opt-in on revert/defect visibility, point maintainer process at the canonical GitHub file, and avoid nested H1/H2 inside the August archive snapshot. Assisted-by: Cursor Grok 4.6 Signed-off-by: Adam Scerra <ascerra@redhat.com> Co-authored-by: Cursor <cursoragent@cursor.com>
|
🤖 Finished Review · ✅ Success · Started 2:38 AM UTC · Completed 2:54 AM UTC Commit: Runtime: claude · Model: opus → claude-opus-4-6 · Effort: high · Cost: $5.46 |
|
Risk Assessment: moderate (2/5) DetailsDocs-focused roadmap refresh with trivial 1-line touches to Makefile and pre-commit config; elevated slightly by CI workflow flag and protected path hit, but all non-doc changes are minimal config adjustments with no functional impact. Previous runRisk Assessment: moderate (2/5) DetailsDocs-focused roadmap refresh with trivial 1-line touches to Makefile and pre-commit config; elevated slightly by CI workflow flag and protected path hit, but all non-doc changes are minimal config adjustments with no functional impact. Previous run (2)Risk Assessment: moderate (2/5) DetailsDocs-focused roadmap refresh with trivial 1-line touches to Makefile and pre-commit config; elevated slightly by CI workflow flag and protected path hit, but all non-doc changes are minimal config adjustments with no functional impact. Previous run (3)Risk Assessment: moderate (2/5) DetailsDocs-focused roadmap refresh with trivial 1-line touches to Makefile and pre-commit config; elevated slightly by CI workflow flag and protected path hit, but all non-doc changes are minimal config adjustments with no functional impact. |
ReviewFindingsHigh
Low
Next steps:
Previous runReviewFindingsHigh
Low
Next steps:
Previous run (2)ReviewFindingsHigh
Low
Next steps:
Previous run (3)ReviewFindingsHigh
Low
Next steps:
|
![fullsend-ai-review[bot]](https://avatars.githubusercontent.com/in/3478501?s=60&v=4){kind=small}
This comment was marked as outdated.
This comment was marked as outdated.
Sorry, something went wrong.
August was only on /archived-roadmaps while the sidebar still opened the old combined file, so August never appeared. Put July, June, and Foundation next to August as dated files under archived-roadmaps/. Assisted-by: Cursor Grok 4.6 Signed-off-by: Adam Scerra <ascerra@redhat.com> Co-authored-by: Cursor <cursoragent@cursor.com>
|
🤖 Finished Review · ✅ Success · Started 10:56 AM UTC · Completed 11:11 AM UTC Commit: Runtime: claude · Model: opus → claude-opus-4-6 · Effort: high · Cost: $4.84 |
This comment was marked as outdated.
This comment was marked as outdated.
Sorry, something went wrong.
Bring the September roadmap branch onto current main. Assisted-by: Cursor Grok 4.6 Signed-off-by: Adam Scerra <ascerra@redhat.com> Co-authored-by: Cursor <cursoragent@cursor.com>
Call out ADR 0044 for leftover per-org cleanup, and drop unrelated prettier-only VitePress config churn. Assisted-by: Cursor Grok 4.6 Signed-off-by: Adam Scerra <ascerra@redhat.com> Co-authored-by: Cursor <cursoragent@cursor.com>
|
🤖 Finished Review · ✅ Success · Started 11:19 AM UTC · Completed 11:36 AM UTC Commit: Runtime: claude · Model: opus → claude-opus-4-6 · Effort: high · Cost: $7.13 |
This comment was marked as outdated.
This comment was marked as outdated.
Sorry, something went wrong.
waynesun09
left a comment
waynesun09
left a comment
There was a problem hiding this comment.
One documentation-accuracy finding inline on docs/roadmap.md.
Give standing user-support work its own last-place theme, and label August Jira leftovers as issues vs ADR PRs per review. Assisted-by: Cursor Grok 4.6 Signed-off-by: Adam Scerra <ascerra@redhat.com> Co-authored-by: Cursor <cursoragent@cursor.com>
|
🤖 Finished Review · ✅ Success · Started 8:22 PM UTC · Completed 8:36 PM UTC Commit: Runtime: claude · Model: opus → claude-opus-4-6 · Effort: high · Cost: $4.57 |
| - id: lint-md-links | ||
| name: lint markdown links | ||
| entry: lychee --offline --no-progress --include-fragments --exclude-path node_modules --exclude-path experiments --exclude-path docs/archived-roadmap.md | ||
| entry: lychee --offline --no-progress --include-fragments --exclude-path node_modules --exclude-path experiments --exclude-path docs/archived-roadmaps |
There was a problem hiding this comment.
[high] protected-path
PR modifies a protected governance/infrastructure file. The change updates lychee --exclude-path from docs/archived-roadmap.md to docs/archived-roadmaps to match the new archive directory structure. The PR body acknowledges this requires human approval, but no linked issue provides authorization. Human sign-off is required for protected-path changes.
Suggested fix: Human maintainer must approve the .pre-commit-config.yaml change. The exclude-path update is mechanically required by the archive restructuring, but the file protected status demands explicit sign-off.
| @@ -314,9 +314,7 @@ export default defineConfig({ | |||
| { | |||
There was a problem hiding this comment.
[low] scope-coherence
Two formatting-only changes unrelated to the roadmap refresh: collapsing the Reference items array to a single line and reformatting the OpenAI Workload Identity entry to multi-line. These appear to be auto-formatter output.
| @@ -1,732 +1,13 @@ | |||
| # Archived roadmaps | |||
There was a problem hiding this comment.
[low] architectural-coherence
Old monolithic archive file gutted to redirect stub pointing to new per-month files under docs/archived-roadmaps/. Existing bookmarks continue to resolve. Sidebar config uses getMarkdownFiles with .reverse(), aligning with the established auto-discovery pattern.
waynesun09
left a comment
waynesun09
left a comment
There was a problem hiding this comment.
Automated review sweep — inline findings below, plus one finding on an unchanged line the diff doesn't cover:
docs/.vitepress/config.ts (line 442, outside diff hunks)
[MEDIUM] archived-roadmaps folder is entirely absent from VitePress search scopes (overcorrection from an earlier review round)
Verified via git history on this PR: commit 34086ff (the PR's first commit) added "/docs/archived-roadmaps/" to the Contributing scope's prefixes array, which drew a qodo-code-review[bot] comment (id 3910157431, orig_line 442) calling that a misclassification (archived pages showing up under the live "Contributing" filter). A later commit, cd5ecbbb, responded by removing "/docs/archived-roadmaps/" from the Contributing prefixes entirely rather than moving it to its own scope. On the current head (22cab23), search.options.scopes lists Guides, Design Docs, Experiments, Contributing (["/docs/contributing/"] only), and a catch-all Others — /docs/archived-roadmaps/ appears in no scope's prefix list. The archived pages are still placed under "Contributing" in the sidebar navigation (config.ts ~line 368-370), so with the Contributing search pill active, the archived roadmap pages are now unreachable there and only surface under the generic "Others" scope — inconsistent with their sidebar placement. This is a different, currently-live defect from the one the earlier bot comment flagged (which described the opposite state and was already addressed by the intervening commit), so no existing comment covers it. Note: line 442 (the unchanged scopes array) falls outside this PR's diff hunks, so this finding is included in the review body rather than as an inline comment.
Suggestion: Add "/docs/archived-roadmaps/" to an appropriate scopes entry — either a new dedicated scope (e.g. "Archived Roadmaps") or back into Contributing's prefixes if that classification is intended — so local search matches the sidebar placement, without reintroducing the earlier misclassification-under-Contributing complaint if a dedicated scope is used instead.
| Parked or dependency-bound in this planning session. Not dated. | ||
|
|
||
| - Isolated sandboxes per subagent (host APIs come first): [fullsend#3978 — harness: subagents share the parent sandbox instead of running in isolated sandboxes](https://github.com/fullsend-ai/fullsend/issues/3978) | ||
| - Red Hat AI inference routing — waiting on an inference proxy from another team: [fullsend#6464](https://github.com/fullsend-ai/fullsend/issues/6464) |
There was a problem hiding this comment.
[HIGH] Wrong issue linked for "Red Hat AI inference routing" item
Line 151 reads: Red Hat AI inference routing — waiting on an inference proxy from another team: [fullsend#6464](https://github.com/fullsend-ai/fullsend/issues/6464). Verified via gh api repos/fullsend-ai/fullsend/issues/6464 that its actual title is "Track pi (earendil-works/pi) as a supported agent runtime" (state: open) — an unrelated topic. The same issue number is used correctly elsewhere in this PR's own diff: docs/archived-roadmaps/2026-08.md line 82 cites fullsend#6464 for "Pi runtime pilot — Second production runtime; Bedrock/Azure access," confirming #6464 really is the pi-runtime issue, not an inference-routing one. No existing PR comment addresses this line/issue-number mismatch.
Suggestion: Link the correct tracking issue for the Red Hat AI inference-routing/proxy work if one exists, or drop the issue link and present it as an undated direction bullet with no tracking issue (matching the pattern already used for "Partners" elsewhere in this file).
| @@ -1,732 +1,13 @@ | |||
| # Archived roadmaps | |||
There was a problem hiding this comment.
[MEDIUM] Old archive URL is a plain stub, not a canonical redirect (SEO duplicate + broken anchors)
Verified on head commit 22cab23: docs/archived-roadmap.md is now a plain markdown stub with a table of links to the new dated pages, not a redirect. The repo already has an established moved-page pattern for exactly this situation — docs/index.md uses <meta http-equiv="refresh" content="0; url=guides/getting-started/"> and docs/.vitepress/seo.ts's CANONICAL_REDIRECTS maps "index.md": "guides/getting-started/" so crawlers see one canonical URL. CANONICAL_REDIRECTS in the head version of seo.ts contains only the index.md entry — archived-roadmap.md is not mapped, and the stub carries no meta-refresh. This leaves the old /docs/archived-roadmap URL as a second indexable page competing with the new /archived-roadmaps/ index, doesn't send users/crawlers to the new canonical location, and silently drops old in-page fragment anchors that may be linked from elsewhere. This is distinct from the existing low-severity fullsend-ai-review[bot] comment on this same file/line (id 3918420404), which only notes that old bookmarks still resolve via the stub table and does not raise the canonical-redirect/duplicate-indexing concern.
Suggestion: Add archived-roadmap.md to CANONICAL_REDIRECTS in docs/.vitepress/seo.ts (mapping to archived-roadmaps/) and add the same meta-refresh used in docs/index.md, consistent with the site's existing moved-page convention; consider noindex on the stub if it remains reachable.
| - id: lint-md-links | ||
| name: lint markdown links | ||
| entry: lychee --offline --no-progress --include-fragments --exclude-path node_modules --exclude-path experiments --exclude-path docs/archived-roadmap.md | ||
| entry: lychee --offline --no-progress --include-fragments --exclude-path node_modules --exclude-path experiments --exclude-path docs/archived-roadmaps |
There was a problem hiding this comment.
[MEDIUM] lychee link-check exclude widened from one file to an entire directory, hiding newly-added/rewritten links from checking
Verified: on origin/main the lint-md-links hook excludes a single file (--exclude-path docs/archived-roadmap.md); on head commit 22cab23 it excludes the whole directory (--exclude-path docs/archived-roadmaps). The original single-file exclude existed because that file held historical #fragment-style headings inside <details> blocks that aren't real anchors. This PR's directory-wide exclude also blinds lychee to the live relative links this same PR introduces inside docs/archived-roadmaps/ (e.g. ../roadmap.md, ../problems/*.md, ../ADRs/0028-gitlab-support.md links in the dated pages) and the new docs/archived-roadmaps/README.md index page — none of which will ever be link-checked going forward. This is separate from the existing HIGH "protected-path" bot comments on the same line (ids 3913467857/3913649038/3918420382), which only flag that .pre-commit-config.yaml is a protected file requiring human sign-off and say nothing about the exclude scope hiding real links from the checker.
Suggestion: Narrow the exclude to just the fragment-heavy content that needs it (e.g. a path/pattern targeting only the historical <details> fragment sections) rather than excluding the whole docs/archived-roadmaps directory, so relative links in README.md and the dated pages stay covered by lychee.
2 participants
Summary
docs/archived-roadmaps/(2026-08.md,2026-07.md,2026-06.md,2026-04.md) plus an index.docs/archived-roadmap.mdis a stub so old bookmarks keep working.Deprecated per-org (ADR 0044)
This PR mentions leftover cleanup from deprecated per-org installation mode (ADR 0044), tracked as fullsend#2887. That is remaining teardown, not new org-mode architecture.
Protected path
.pre-commit-config.yamlupdates lychee--exclude-pathfrom the old single archive file todocs/archived-roadmapsso monthly archives are not link-checked. Needs human approval; the exclude is required for the new layout.Test plan
archived-roadmaps/index and dated pages;/archived-roadmapstub still resolves.2046baf8was harnesspr-pingflake onhalfsend-02/test-repo-02, not this docs change; re-run after this push.Made with Cursor