Release-note fact check — anthropics/claude-code

The only evidence is the feed.xml changelog entry that restates the claim itself; per the rules, a changelog hunk cannot serve as evidence for its own claims. No code changes implementing the background subagent behavior for /code-review are shown in the diff.

Identifiers Option+Delete, Ctrl+W, Cmd+Backspace, Ctrl+U, Ctrl+K all appear in the diff (1 file(s)).

Identifiers \u, C:\Users\unicorn all appear in the diff (1 file(s)).

The evidence shows the claim restated verbatim in the release notes feed.xml, but the rules explicitly state: 'A changelog or release-notes hunk restating the claim is NOT evidence — notes cannot prove themselves. A claim about code behavior needs code changes as support.' No code diffs implementing the left-arrow/Esc handling changes are provided, only the documentation restatement.

Identifiers claude mcp list, /mcp all appear in the diff (1 file(s)).

The evidence shows only feed.xml with release notes text restating the claim, which per the rules is not evidence (changelog hunks cannot prove themselves). No actual code changes are provided in the diff to demonstrate the underlying fix; docs hunks only support claims about documentation, not code behavior.

The diff shows only the release notes (feed.xml) restating the claim. Per the rules, a changelog hunk cannot prove itself — a claim about code behavior requires code changes as support, not documentation hunks restating the claim.

The diff shows the claim restated verbatim in feed.xml release notes, but release-notes hunks cannot prove themselves per the rules. No linked commits are provided and the only shown files are feed.xml and CHANGELOG.md (release documentation), not code changes demonstrating the fix was implemented.

The diff shows only documentation changes (feed.xml and CHANGELOG.md entries) restating the claim. Per the rules, a changelog hunk restating a claim is not evidence; a claim about code behavior requires code changes as support. No linked commits or implementation changes are provided.

The diff shows only a feed.xml changelog entry restating the claim verbatim, which cannot serve as evidence per the rules ('docs hunks only support claims about documentation'). No code changes demonstrating the gateway spend metering fix are provided; CHANGELOG.md is listed as changed but its diff is not shown.

The evidence contains only a changelog/release-notes hunk (feed.xml) restating the claim verbatim. Per the rules, changelog entries cannot serve as evidence for code behavior claims — code changes are required. No commits are linked and the file list shows only documentation files changed, with no actual implementation diffs provided.

The diff contains only documentation changes—changelog and feed entries that restate the claim verbatim. Per the explicit rule, 'a changelog or release-notes hunk restating the claim is NOT evidence'; a claim about code behavior requires code changes as support. No linked commits, no actual source implementation, no supporting diffs.

The feed.xml diff restates the claim verbatim in release-notes format, but per the rules, release notes cannot prove themselves — code changes are needed to verify a behavioral fix. Only documentation files (feed.xml, and CHANGELOG.md not shown) were provided; no implementation diffs are present to support the claim about fixing spurious messages or unpaired tool_use blocks.

Identifiers --ax-screen-reader, VoiceOver all appear in the diff (1 file(s)).

The evidence shows only feed.xml and CHANGELOG.md (documentation files) being changed. The diff restates the claim verbatim in the v2.1.218 release notes, but per the stated rules, 'A changelog or release-notes hunk restating the claim is NOT evidence.' A claim about code behavior (fixing terminal cursor movement) requires code changes to support it, not a changelog entry that merely repeats the claim.

The diff shows only a release notes entry in feed.xml restating the claim verbatim; this is a changelog restatement, not evidence of code changes implementing the fix. Per the rules, 'A changelog or release-notes hunk restating the claim is NOT evidence — notes cannot prove themselves.' No actual code changes are shown that would support the crash fix.

The feed.xml diff shows the release notes entry restating the exact claim, but per the rules, 'A changelog or release-notes hunk restating the claim is NOT evidence.' No code changes are provided (linked commits: none; only feed.xml and CHANGELOG.md shown), so there is no evidence that the bug fix was actually implemented.

The diff shows only feed.xml (release notes) being updated with the claim text restated. Per the rules, 'A changelog or release-notes hunk restating the claim is NOT evidence — notes cannot prove themselves.' Code behavior claims require code changes as support; no actual implementation diffs are provided.

The evidence shows the claim appearing verbatim in the release notes (feed.xml), but the rules explicitly state that a changelog or release-notes hunk restating the claim is NOT evidence. A claim about code behavior (using a monotonic clock for timing) requires actual code changes as support; no implementation files are shown in the provided diffs, only documentation metadata.

The diff shows only feed.xml (release notes) containing the claim text verbatim; per the rules, a changelog restating the claim is not evidence. No code changes, CHANGELOG.md diff, or implementation details are provided to verify the fix actually exists.

The diff shows only a release-notes entry in feed.xml restating the claim verbatim. Per the review rules, changelog entries cannot serve as evidence for code behavior claims — actual implementation changes are required. No code diffs or commit evidence supporting the fix are provided.

The diff shows only a changelog entry in feed.xml that restates the claim verbatim ('Fixed a retry loop that re-sent identical doomed requests after a context-overflow error with a large thinking budget; Ctrl+B backgrounding now applies the same background-shell caps as other paths'). Per the rules, 'A changelog or release-notes hunk restating the claim is NOT evidence — notes cannot prove themselves. A claim about code behavior needs code changes as support.' No actual implementation changes ar

The evidence contains only the changelog entry restating the claim itself (in feed.xml), which per the rules cannot prove a code behavior change. No diffs to actual implementation code are provided, and the referenced CHANGELOG.md change is not shown in detail.

The diff shows only feed.xml and CHANGELOG.md changed, both documentation. The feed.xml entry merely restates the claim; per instructions, a changelog hunk restating a claim is not evidence—a code behavior fix requires code changes as support, not a note about itself.

The diff shows only feed.xml (a release-notes document) being updated with a changelog entry that restates the claim verbatim. Per the rules, 'A changelog or release-notes hunk restating the claim is NOT evidence — notes cannot prove themselves.' No code changes are provided showing the actual fix to session resume logic or delta-attachment handling.

The diff shows only feed.xml with the release-notes entry restating the claim verbatim ('Improved /ultrareview error feedback so Claude can correct an invalid argument instead of retrying it unchanged'). Per the rules, a changelog hunk cannot prove itself—code behavior claims require code changes as support. No actual implementation changes are shown in the provided evidence.

The feed.xml diff contains an HTML-encoded release note that restates the claim verbatim, but per the rules, 'A changelog or release-notes hunk restating the claim is NOT evidence' — notes cannot prove themselves. A behavioral claim about permission-dialog handling requires actual code changes, not documentation that describes the change. No code diffs or CHANGELOG.md details are provided to verify the implementation.

The only evidence is the release notes entry in feed.xml that restates the claim verbatim ('Improved sandbox command restrictions for IDE interactions'). Per the rules, changelog hunks cannot prove themselves; code changes are required to support claims about behavior. No actual implementation changes are provided in the diff to show what restrictions were modified or how.

The only diff provided is feed.xml, which is release-notes documentation restating the claim itself. Per the rules, 'A changelog or release-notes hunk restating the claim is NOT evidence — notes cannot prove themselves.' CHANGELOG.md is listed as changed but no diff is shown. No code changes demonstrating the actual implementation of the trust-dialog improvement are provided.

The evidence is a release-note entry in feed.xml that restates the claim verbatim. Per the rules, a changelog restating the claim is not evidence—behavior claims require code changes, not documentation hunks. No implementation diff is provided.

The diff shows only feed.xml (release notes) being updated with text that restates the claim verbatim. Per the rules, a changelog or release-notes hunk restating the claim is NOT evidence — actual code changes showing the auto-mode classifier implementation are needed to verify a behavioral change claim.

The only evidence shown is the release notes in feed.xml, which restates the claim verbatim. Per the rules, changelog entries restating a claim are not evidence—a claim about code behavior requires code changes as support. No source files or implementation diffs are provided.

The only diff provided is feed.xml, which is the release notes feed. Per the evaluation rules, a changelog or release-notes hunk restating the claim is NOT evidence—notes cannot prove themselves. The claim requires code changes showing how server-managed settings behavior changed; no such implementation diffs are provided.

The diff shows only documentation files (feed.xml and CHANGELOG.md) containing a release-note entry that restates the claim verbatim. Per the rules, changelog hunks cannot prove themselves — the claim requires code changes (agent file validation logic) as evidence, not documentation that describes it. No implementation files are shown in the diff.

Identifiers context: fork, background: false all appear in the diff (1 file(s)).

Identifiers yes, no, on, off, 1 all appear in the diff (1 file(s)).

The diff shows the claim text appearing verbatim in feed.xml (release notes), but per the fact-check rules, 'a changelog or release-notes hunk restating the claim is NOT evidence — notes cannot prove themselves.' The provided evidence contains only documentation changes (feed.xml, CHANGELOG.md); no code diffs are shown to demonstrate the actual fix was implemented.