Skip to content

fix(chat): show only the host on a blocked-link chip - #20267

Merged
bolichen97 merged 1 commit into
mainfrom
fix/blocked-link-chip-domain-only-18673
Oct 14, 2026
Merged

bolichen97 merged 1 commit into
mainfrom
fix/blocked-link-chip-domain-only-18673

Conversation

@bolichen97

@bolichen97 bolichen97 commented Oct 14, 2026 •

Copy link
Copy Markdown
Contributor

Problem / Motivation

Goal: A Blocked link chip names only what it can prove about its own link: the site name.

The chip prints a record's path and query length beside the host whenever the host has exactly one blocked-link record in the message. One record does not prove it is this chip's link. A presigned S3 link is split at &X-Amz-Date, so its placeholder has no record of its own; a second same-host URL that a credential pass redacted does have a record (/other-a). Where the presigned link was, the chip reads my-bucket.s3.us-east-1.amazonaws.com/other-a ⋯ query 41 chars. Markdown such as [REDACTED: suspicious URL to h.example-sink.net] renders to a chip no record stands behind, and it reads h.example-sink.net/real ⋯ query 290 chars like the real one.

Why it matters

The reader is told the removed link went to an address it never went to, and the details card then appears to confirm it. A chip that can name the wrong address is worse than one that names only the host, because the whole point of the chip is to tell the reader what was removed.

Not a goal

  • Pairing placeholders with records. That order cannot be made sound on the production redaction pipeline.
  • Changing what the details card shows. It already lists each record under its own address.
  • Changing which records the backend keeps or what path / query_chars hold.

What changed (motivation → approach → change)

The chip reads the host's record group and, when the group has one record, prints that record's path and query_chars. Nothing ties a chip to a record: the saved placeholder carries only the host, a placeholder can have no record, and the rendered text can hold more chips than placeholders. So "one record" never means "this chip's record".

BlockedLinkChip in website/src/components/RedactionCards.tsx drops the single-record branch and prints the host alone. The card is unchanged: a lone record keeps main's shape (Destination host, reason, the kept full address or the withheld note), and several records are listed each under its own address, where no entry is presented as the link clicked. A lone record's path is therefore shown nowhere, on purpose: under a card titled "This link is blocked" it would make the same unprovable claim the chip made. The components.redaction.link_query string has no caller left, so it leaves every catalog under website/src/i18n/locales/ (en-XA is regenerated by gen-pseudolocale.mjs). The chip rule in docs/request-for-change/rfc-redaction-explain-and-reveal.md §3 says what the chip may show and why, and docs/system-specs/modules/security.md names the host-only label. website/scripts/capture-blocked-link-chip-host-only.mjs shoots the two cases from the issue against the built dashboard and fails if any chip shows a path or query length.

Case Before After
one placeholder, one record of another address host/other-a ⋯ query 41 chars host
forged chip beside one real record host/real ⋯ query 290 chars on both host on both
several records, same host host host

The presigned-link tail (&X-Amz-Date=…) that runs on as plain text beside the chip comes from the redaction split itself and is on main today; this PR leaves it as it is.

Backwards compatibility

Compatible: the chip is display only; no API, record shape, or stored message changes. The card still shows every record's path, reason and full address.

Reader: website/src/components/RedactionCards.tsx:BlockedLinkChip -- chat -- the only consumer of components.redaction.link_query; the branch that read it is gone, and website/src/i18n/deadKeys.test.ts holds its baseline because the key leaves with its caller.

Tests

  • website/src/test/MarkdownRenderer.blockedLink.test.tsx: the one-record chip shows the host alone and no blocked-link-query; a twice-recorded address and two differing records both label the chip with the host.
  • website/src/test/MarkdownRenderer.blockedLinkOneOne.test.tsx: the blockedLinkOneOne.json case (one placeholder, one record of another address) names the host and not /other-a on the chip; its card names the host and never /other-a as the link, with no full-URL view.
  • website/src/test/MarkdownRenderer.blockedLinkForgedChip.test.tsx: a forged chip and the real chip beside it both show the host alone, with no path or query length, before each opens the honest host list.
  • With RedactionCards.tsx at main, six of these fail; with the change, all pass.

Manual verification

Built the dashboard and ran node website/scripts/capture-blocked-link-chip-host-only.mjs: three chips, every target equals its host, no query-length span, and the presigned-link chip's card names the host, never /other-a, with no full-URL view.

Screenshots / video

Before (main): the presigned-link chip names /other-a and 41 chars; both h.example-sink.net chips name /real and 290 chars.

before, dark

After: every chip names its host alone; the forged chip's card explains the one real record of h.example-sink.net as on main.

after, dark

After, the presigned-link chip's card: the lone record's full address is withheld; the card names the host and the reason and nowhere says /other-a is the removed link.

after, withheld record card, dark

Light theme

before, light

after, light

after, withheld record card, light

Related Issues

Closes #18673

Pattern harvest

Rule candidate: review-prompt
Pattern: a UI label built from a lookup that returns one item treats "exactly one candidate" as "the matching one" without a key that proves the match.

Checklist

  • At most two commits (one is the norm), with a Conventional Commits title (feat|fix|docs|style|refactor|perf|test|chore|ci|build|revert: ...)
  • Existing tests pass and new tests added for new functionality
  • Self-review completed; code follows project style guidelines
  • Documentation updated (if applicable)
  • No secrets, credentials, or internal references in the diff

@bolichen97
bolichen97 requested review from a team and KiroCrewCaptain as code owners October 14, 2026 11:21
@bolichen97
bolichen97 requested a review from mbriones98 October 14, 2026 11:21
@bolichen97
bolichen97 enabled auto-merge (squash) October 14, 2026 11:21
@github-actions github-actions Bot added the readiness: checking Automated validation is still running label Oct 14, 2026
@github-actions

github-actions Bot commented Oct 14, 2026 •

Copy link
Copy Markdown
Contributor

First Principles Review (Opus 5.5) — ✅ PASS

Premise-level review of 8fbe4a8e4127c2b26aba6c571c53d97e819dcd16 — why this exists and whether the shipped surface is the smallest honest version. Updated in place on each push. A BLOCK verdict blocks PR readiness; PASS/CONCERNS are advisory.

First-Principles-Verdict: PASS

This PR rewrites the chip rule in RFC §3, whose base status is partial. A maintainer should confirm that a chip showing the host alone is the agreed shape.

What this change ships

Inventory (7 items) — 7 justified

Intent: stop a Blocked link chip from claiming a removed link went to an address it never went to. This is a FIX (#18673).

  1. When a host has one record, the chip no longer shows that record's path — justified
  2. The chip no longer shows "⋯ query N chars" — justified
  3. When the same address is recorded twice, the chip shows the host only, not host plus path — justified
  4. The components.redaction.link_query string is removed from all 14 locale catalogs; it has no callers left — justified
  5. The RFC §3 chip rule now says the chip shows the site name only — justified
  6. The security.md bullet now says the chip shows the host only — justified
  7. A new screenshot script, capture-blocked-link-chip-host-only.mjs, follows the repo's 746 existing capture scripts — justified

[FIRST-PRINCIPLES-REVIEWED] 8fbe4a8

Description read

[DESCRIPTION-READ] 6aad116c4698deae950cbb504295421aeddf8aad5f8309c4fcdd9067bf9020d1

That sha256 names the PR description this verdict read. Recompute it with .github/scripts/pr-description-capture.sh; if it differs, the description changed after this verdict was formed and any finding drawn from the description is unproven.

@github-actions

github-actions Bot commented Oct 14, 2026 •

Copy link
Copy Markdown
Contributor

Design Review (Opus 5.5) — 🟡 CONCERNS

Design-level review of 8fbe4a8e4127c2b26aba6c571c53d97e819dcd16 — updated in place on each push. A BLOCK verdict blocks PR readiness; PASS/CONCERNS are advisory.

Design-Verdict: CONCERNS

The card still makes the claim the chip gave up: a lone record with a kept URL keeps showing its full address and "Open once" under any chip of that host.

Watch

  • The card still treats the one record as this chip's link. The PR's own premise is that "one record never means 'this chip's record'". The chip now follows that premise, but BlockedLinkEntry doesn't. For a lone record whose url was kept, the card still renders blocked-link-review (the full URL), blocked-link-copy and blocked-link-open-once under "This link is blocked". The description confirms this for the forged chip: "the forged chip's card explains the one real record of h.example-sink.net as on main". The chain: a forged chip, or a presigned placeholder with no record of its own, opens the card. The card offers to open or copy a different URL. The reader acts on an address the removed link never had, which is the harm the PR was opened to remove, now one click deeper. The OneOne case only avoids this because that record's URL happens to be withheld. The description also says "A lone record's path is therefore shown nowhere". That is untrue whenever the URL is kept, because the full URL contains the path.
    Clears when: either the lone-record card frames its record as "a blocked link to this host" (the same list-style framing the several-record case uses, with no claim that it is this link), or the PR states that "Changing what the details card shows" remains a known gap and links a follow-up issue. The PR should also fix the "shown nowhere" sentence.

[DESIGN-REVIEWED] 8fbe4a8

Description read

[DESCRIPTION-READ] 7187a3197c4039c04235e1fb3891ac0d7a846078bfad19a5bc13f633d126805f

That sha256 names everything this verdict read: the PR description, plus the 1 evidence file(s) the workflow collected from it and handed this review. It covers that evidence because the media strip replaces every attachment URL with the same placeholder, so an edit swapping one attachment for another would leave the description's captured text identical while what this review judged changed. Recompute it with .github/scripts/pr-description-capture.sh over the same inputs; if it differs, the description or that evidence changed after this verdict was formed and any finding drawn from either is unproven.

@github-actions

github-actions Bot commented Oct 14, 2026 •

Copy link
Copy Markdown
Contributor

UX Review (Opus 5.5) — ✅ PASS

UX-level review of 8fbe4a8e4127c2b26aba6c571c53d97e819dcd16 — updated in place on each push. A BLOCK verdict blocks PR readiness; PASS/CONCERNS are advisory.

UX-Verdict: PASS

Check whether the plain-text &X-Amz-Date=… tail beside the chip needs a follow-up. It is on main, but the blind reader said it "makes the shortened pill feel pointless".

[UX-REVIEWED] 8fbe4a8

Description read

[DESCRIPTION-READ] 1cc73cb68d9b59d2639c8c5aa92d36876213a90e64927109e33445e68b2b8aba

That sha256 names everything this verdict read: the PR description, plus the 9 evidence file(s) the workflow collected from it and handed this review. It covers that evidence because the media strip replaces every attachment URL with the same placeholder, so an edit swapping one attachment for another would leave the description's captured text identical while what this review judged changed. Recompute it with .github/scripts/pr-description-capture.sh over the same inputs; if it differs, the description or that evidence changed after this verdict was formed and any finding drawn from either is unproven.

@github-actions

github-actions Bot commented Oct 14, 2026 •

Copy link
Copy Markdown
Contributor

Opus 5.5 Review — ✅ no blocking findings

Reviewed 8fbe4a8e4127c2b26aba6c571c53d97e819dcd16 — this comment is updated in place on each push.

Review details

No findings.

[OPUS-REVIEWED] 8fbe4a8

Verdict parsed from the review's SHA-scoped output markers for commit 8fbe4a8e4127c2b26aba6c571c53d97e819dcd16.

False positive or not applicable? A repository writer can comment:
/ai-review override fable 8fbe4a8e4127c2b26aba6c571c53d97e819dcd16: <one-sentence reason>

@github-actions

github-actions Bot commented Oct 14, 2026 •

Copy link
Copy Markdown
Contributor

GPT 6.1 Review — ✅ no blocking findings

GPT 6.1 completed its review of 8fbe4a8e4127c2b26aba6c571c53d97e819dcd16 and found no blocking issues.

This comment is updated in place on each push.

Review details

No findings.
[GPT-REVIEWED] 8fbe4a8

False positive or not applicable? A repository writer can comment:
/ai-review override gpt 8fbe4a8e4127c2b26aba6c571c53d97e819dcd16: <one-sentence reason>

@bolichen97

Copy link
Copy Markdown
Contributor Author
  • Item 1 — product shape, decided by this PR alone — needs-a-decision (span=32ff8470face)

needs-a-decision: a maintainer is asked to confirm or reject the RFC §3 chip rule in this diff ("the chip shows only what is proven about that chip, which is the site name"), either in the approval or with /ai-review override first-principles ef67f150998d6fa105fdffa7b493c7e8ca9482bf: <decision>.
What the base rule allowed: blockedLinkOneOne.json (backend output, pinned by test_chat_runner_coverage.py) renders a chip that names /other-a ⋯ query 41 chars for a link that never went there; a forged &#91;…&#93; chip names /real for a link no record stands behind. Both fail on main; #18673 reports them.
Why the rule, not just the code: the saved placeholder carries only the host, a placeholder can have no record, and the rendered text can hold more chips than placeholders, so no record count proves which record a chip stands for. The same conclusion was reached and left open on #17613 (its disposition "Item 2: agreed, needs a maintainer").
The agent does not post this override: the finding asks for a product ruling, which is a person's call.

[operator: bolichen97]

@bolichen97

Copy link
Copy Markdown
Contributor Author
  • Check that the RFC owner accepts dropping the path and query length from the chip — needs-a-decision

needs-a-decision: the same ask as First Principles Item 1; the RFC owner (rayrayxu per the RFC header) or an approving maintainer is asked to confirm the §3 chip rule on this head.
The old test's "matches the prototype" named the layout in docs/request-for-change/assets/redaction-explain-reveal-prototype.html, which drew the chip before the backend showed that a one-record host can be the wrong record; the test now names what the chip can prove.

[operator: bolichen97]

@bolichen97

Copy link
Copy Markdown
Contributor Author
  • The leftover URL text now looks like spill from the chip — accepted-and-deferred (span=ba92479c17e5)

accepted-and-deferred: the tail comes from the redactor splitting a presigned link at &X-Amz-Date (the delta path, TestBlockedLinkCardRecords._one_presigned_one_key_split_flush), which is on main and outside this PR's goal of what the chip may claim; the fix sits on the redaction or renderer side, not in the chip label.
Tracked in #20279 (label deferred-finding, assignee bolichen97, Due 2026-10-28). The PR body names the tail as left as it is.

[operator: bolichen97]

@bolichen97

Copy link
Copy Markdown
Contributor Author
  • The two h.example-sink.net chips look identical, and so does their card — rebutted (span=31a5c2818020)

rebutted: identical chips are the correct shape here, as the lane notes: nothing the reply carries ties a chip to a record, so two chips of one host cannot be told apart without inventing a link; ## Not a goal names pairing placeholders with records.
The card heading components.redaction.link_title ("This link is blocked…") is main's wording, kept on purpose (## Not a goal: changing what the card shows); when the host has several records the card lists each under its own Destination: line, so the reader sees all of them. Rewording the heading for the several-records case is a card copy change for #17613's follow-up line, not this fix.

[operator: bolichen97]

@bolichen97

Copy link
Copy Markdown
Contributor Author
  • Rename Inspect to Why blocked? — rebutted (span=41395c3607c1)

rebutted: the button label components.redaction.link_inspect is unchanged by this PR and shared with the credential lock tag's card opener; renaming it is a copy decision across both redaction surfaces and 13 catalogs, outside the goal of what the chip label may claim.
The chip is not itself clickable (it is a role=group span, never an anchor), so the button is the only opener; its label is a separate UX item for the RFC owner, not a defect in this diff.

[operator: bolichen97]

@bolichen97

Copy link
Copy Markdown
Contributor Author
  • A lone withheld record's path shows nowhere once the chip names the host alone — fixed in ef67f15 (span=9a87a2fb0203)

fixed: BlockedLinkCard passes showTarget={group.length > 1 || record.path != null}, so every record that carries a path is listed under its own address in the card, a single withheld one included.
Pinned by MarkdownRenderer.blockedLinkOneOne.test.tsx ("the card names the lone withheld record by its path"): the entry reads my-bucket.s3.us-east-1.amazonaws.com/other-a and no full-URL view exists; blockedLinkCardAsserts.ts expects the address line for every record with a path.

[operator: bolichen97]

@bolichen97
bolichen97 force-pushed the fix/blocked-link-chip-domain-only-18673 branch 2 times, most recently from 2663eab to 357f8bf Compare October 14, 2026 12:29
The chip named a record's path and query length whenever its host had
exactly one record in the message. One record does not prove it is this
chip's link: a placeholder can have no record of its own (a presigned link
split at its query), and the rendered text can hold more chips than the
saved text has placeholders (an entity-encoded or escaped bracket). So the
chip could name a different address than the one it replaced, and the card
then appeared to confirm it.

The chip now shows the host alone. The card still lists each record under
its own address, reason and actions. The unused `link_query` string leaves
the catalogs, and the RFC §3 chip rule says what the chip may show and why.

Closes #18673
@bolichen97
bolichen97 force-pushed the fix/blocked-link-chip-domain-only-18673 branch from 357f8bf to 8fbe4a8 Compare October 14, 2026 12:53
@bolichen97

Copy link
Copy Markdown
Contributor Author
  • Item 2 — oversized: a lone record's address appears on the card — fixed in 8fbe4a8 (span=d09e61a619c7)

fixed: the mechanism is removed. BlockedLinkCard is back to main's showTarget={group.length > 1}; the card is byte-identical to main for one record and for several, which is what the frozen Not a goal says.
A lone record's path is now shown nowhere, on purpose: under a card titled "This link is blocked" it would make the same unprovable claim the chip made (the UX lane read it exactly that way on 2663eab). The RFC §3 text and blockedLinkOneOne.test.tsx pin that the card never names /other-a as the link.

[operator: bolichen97]

@bolichen97

Copy link
Copy Markdown
Contributor Author
  • Item 3 — oversized: the labelTarget prop and the link_entry_on_host key — fixed in 8fbe4a8 (span=d843b5847ad1)

fixed: the subtraction is taken. labelTarget, the blocked-link-entry-label span and components.redaction.link_entry_on_host are deleted from the component and from every catalog; blockedLinkCardAsserts.ts is restored to main.
The diff now touches the card not at all: the chip loses its single-record branch, the dead link_query key leaves the catalogs, the RFC §3 chip rule and the security spec sentence say so, and the tests and capture script pin it.

[operator: bolichen97]

@bolichen97

Copy link
Copy Markdown
Contributor Author
  • The new address line uses an unfamiliar word and repeats the site name — fixed in 8fbe4a8 (span=7d3bc738520a)

fixed: the line is gone. The card keeps main's lone-record shape (Destination: host, reason, the kept address or the withheld note), so the site name appears once and no card copy says "host".
The address line was added on 2663eab to answer an Opus finding; the lane's own read on that head (the card "names an address the chip can't prove is its own") is why it is removed rather than reworded.

[operator: bolichen97]

@bolichen97

Copy link
Copy Markdown
Contributor Author
  • The host-only chip was read correctly, but only as a guess — rebutted (span=d1206277418c)

rebutted: the lane rates this Low and says it needs no change; the reader's reading ("the same blocked link, now showing only the site name; Inspect opens the explanation") is what the chip does.
The chip label Blocked link, the Inspect button and the card are main's; this PR only removes a path and a count the chip could not prove, so there is nothing new for the reader to learn.

[operator: bolichen97]

@bolichen97

Copy link
Copy Markdown
Contributor Author
  • Change link_entry_on_host to "Blocked address to this destination:" — fixed in 8fbe4a8 (span=a4ced95333c9)

fixed by removal: the key components.redaction.link_entry_on_host no longer exists in any catalog, and the card shows no address line for a lone record, so there is no copy left to reword.

[operator: bolichen97]

@bolichen97

Copy link
Copy Markdown
Contributor Author
  • Check the After card screenshot: the lone record's labelled address line should be the only place a path shows — fixed in 8fbe4a8

fixed: on this head a lone record's path shows nowhere, not on the chip and not in the card, so no surface names it as the removed link; the re-shot After frames (chip, forged chip's card, presigned-link chip's card) are attached to the description.
Several records are still listed each under its own address, as on main, where the list itself says no entry is the clicked link.
The earlier ask (RFC owner to accept dropping path and query length from the chip, and redaction-explain-reveal-prototype.html:377 still drawing the old chip) stays a maintainer call; the prototype is a design-time artifact the RFC text now supersedes, and this PR does not edit it.

[operator: bolichen97]

@bolichen97

Copy link
Copy Markdown
Contributor Author
  • A lone withheld record's path shows nowhere once the chip names the host alone — rebutted (span=9a87a2fb0203)

rebutted: not a defect of this diff. On main the card already shows no address line for a lone record (showTarget={group.length > 1}); the only surface that showed the path was the chip, and the chip showing it is the bug (#18673: it named /other-a for a link that never went there).
Showing the path in the lone-record card instead was tried on 2663eab and read by the UX lane as the same unprovable claim moved into the card, which the PR's own Why it matters names as the harm; the RFC §3 rule now says the path appears only in the list of several records, where no entry is presented as the link clicked.
The record keeps path in meta; nothing is deleted, and the Destination, reason and withheld note still explain the lone record.

[operator: bolichen97]

@bolichen97

Copy link
Copy Markdown
Contributor Author
  • The card still treats the one record as this chip's link — accepted-and-deferred (span=1095d6a41646)

accepted-and-deferred: the lone-record card is main's and is outside this PR's frozen goal (Not a goal: "Changing what the details card shows"); it stays a known gap. Reframing it is a card copy and layout decision for the RFC owner, and the one in-PR attempt (an address line with a label on 2663eab and 357f8bf) was judged oversized by First Principles and misread by the UX lane's reader, so it is not settled here.
Tracked in #20303 (label deferred-finding, assignee bolichen97, Due 2026-10-28), which names both candidate framings and the invariants they must keep.

[operator: bolichen97]

@bolichen97

Copy link
Copy Markdown
Contributor Author
  • Check whether the plain-text &X-Amz-Date=… tail beside the chip needs a follow-up — accepted-and-deferred

accepted-and-deferred: it does, and it is on main: the redactor splits a presigned link at &X-Amz-Date in the delta path, so the fix is on the redaction or renderer side, not in the chip label this PR changes.
Tracked in #20279 (label deferred-finding, assignee bolichen97, Due 2026-10-28); the PR body names the tail as left as it is.

[operator: bolichen97]

@bolichen97

Copy link
Copy Markdown
Contributor Author
  • A maintainer should confirm that a chip showing the host alone is the agreed shape — needs-a-decision

needs-a-decision: a maintainer is asked to confirm or reject the RFC §3 chip rule in this diff ("the chip shows only what is proven about that chip, which is the site name"), in the approval of this PR.
Why the rule, not just the code: the saved placeholder carries only the host, a placeholder can have no record (blockedLinkOneOne.json, backend output), and the rendered text can hold more chips than placeholders (a forged &#91;…&#93; chip), so no record count proves which record a chip stands for; #17613's disposition left the same question open ("Item 2: agreed, needs a maintainer").
The agent posts no override: a product ruling is a person's call.

[operator: bolichen97]

@github-actions github-actions Bot added readiness: passed Eligible automated validation passed for the current revision and removed readiness: checking Automated validation is still running labels Oct 14, 2026
@iamwhatever

Copy link
Copy Markdown
Contributor

HOLD at 8fbe4a8.

This PR makes a Blocked link chip show only its host (RedactionCards.tsx L1202). A single record never proves it is that chip's link, so this fix is right. The card is byte-identical to main, and every current-head AI item has a disposition.

What blocks approval:

  1. A product decision is open. Your own needs-a-decision disposition asks a maintainer to accept the RFC section 3 rule that a chip shows only the site name. Only a maintainer can clear that.
  2. Human review is required by policy. The PR changes how redacted links are displayed, and it edits security.md and the redaction RFC.
  3. Please link the two deferred follow-ups in the PR description. Blocked-link card still frames a lone record as the clicked chip's link #20303 covers the lone-record card that still offers Open once under any chip of that host. Presigned-link query tail runs on as plain text beside the Blocked link chip #20279 covers the presigned-link tail. Both issues cover their items and carry the deferred-finding label, an assignee and a due date, but they are linked only from comments.

#20279 describes X-Amz-Signature=… showing as plain text beside the chip. That is on main and not caused by this PR, but a maintainer may want to treat it as more than cosmetic.

This clears when a maintainer rules on the chip rule and reviews the change, and the description links #20303 and #20279. CI on this head is green. I read the tests but did not run them.

— crew-pr-reviewer (r100)

@mbriones98 mbriones98 left a comment

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Checks green and PR Readiness passed at 8fbe4a8; non-feature change.

@bolichen97
bolichen97 merged commit 75726a8 into main Oct 14, 2026
104 of 112 checks passed
@bolichen97
bolichen97 deleted the fix/blocked-link-chip-domain-only-18673 branch October 14, 2026 16:55
@github-actions github-actions Bot removed the readiness: passed Eligible automated validation passed for the current revision label Oct 14, 2026
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

Blocked-link chip names a record it cannot prove

3 participants