[dependabot]

⚠️ Dependabot is rebasing this PR ⚠️

Rebasing might not happen immediately, so don't worry if this takes some time.

Note: if you make any changes to this PR yourself, they will take precedence over the rebase.

---

Bumps the rust-dependencies group with 2 updates in the / directory: russh and signal-hook.

Updates russh from 0.65.0 to 0.65.1

Release notes

Sourced from russh's releases.

v0.65.1

Security fixes

GHSA-736f-c93q-9fwx - agent lock status not checked after confirmation wait

Agent lock status is now rechecked after the confirmation await again.

GHSA-9264-jvgf-6w2f - kex_done called before host-key verification

client::Handler::kex_done is now only called after server host key verification succeeds

Fixes

• 20c5a93: fix(keys): keep the comment when loading a public key file (#801) (@​JakeHuneau) #801
  • load_public_key now keeps the comment from an OpenSSH public key file, so key.comment() returns it and to_openssh() round-trips. Fixes #713.
• 3a21566: fix(client): accept non-canonical ECDSA P-521 signature mpints after strict decode fails (#803) (@​gugahartmann) #803
  • Fixes key exchange against servers such as the Huawei S6720, which pad P-521 signature scalars with extra leading zero bytes.
• a6be9ba: fix(cipher): count the MAC tag in the minimum packet length check (#802) (@​JakeHuneau) #802
  • Fixes valid packets being rejected when an EtM MAC is in use. Fixes #787.
• 43a46d6: fix(AgentClient): propagate method errors (#808) (@​mdevolde) #808
  • add_smartcard_key, lock, unlock, remove_identity, remove_smartcard_key and extension now return the agent's failure response as an error. Before, they always returned Ok(()).
• 7c59594: fix(server): let Dropbear through strict kex by accounting for the ignored first_kex_packet_follows guess (#805) (@​NatLee) #805
  • Fixes Dropbear clients failing to connect when strict key exchange is on. Fixes #804.

Dependencies

• d10345d: deps: trim aws-lc-rs features to avoid transitive deps (#809) (@​sepointon) #809
  • Removes the unused ring-compatible API of aws-lc-rs and the old untrusted crate it pulled in.
• 0a8cc28: deps: remove dep on generic-array 0.14 compat (#810) (@​sepointon) #810

New Contributors

• @​sepointon made their first contribution in Eugeny/russh#810
• @​NatLee made their first contribution in Eugeny/russh#805
• @​JakeHuneau made their first contribution in Eugeny/russh#802
• @​gugahartmann made their first contribution in Eugeny/russh#803

Full Changelog: Eugeny/russh@v0.65.0...v0.65.1

Commits

• b33092a v0.65.1
• 20c5a93 fix(keys): keep the comment when loading a public key file (#801)
• 3a21566 fix(client): accept non-canonical ECDSA P-521 signature mpints after strict d...
• a6be9ba fix(cipher): count the MAC tag in the minimum packet length check (#802)
• 43a46d6 fix(AgentClient): propagate method errors (#808)
• 56384e6 add Eric-D as a contributor for code (#799)
• 7c59594 fix(server): let Dropbear through strict kex by accounting for the ignored fi...
• d10345d deps: trim aws-lc-rs features to avoid transitive deps (#809)
• 0a8cc28 deps: remove dep on generic-array 0.14 compat (#810)
• 92562d3 tests: adapt to VerifiedPublicKey after #785 (#807)
• See full diff in compare view

Updates signal-hook from 0.4.5 to 0.4.6

Changelog

Sourced from signal-hook's changelog.

0.4.6

• Allow registering 'static flags in addition to Arc ones (#215, #180).

Commits

• 0a1de6b Version 0.4.6
• ae4ece5 Merge pull request #215 from Frost-54/static-atomic-unsafe-trait
• 48b56d5 Deps updates
• fd51cde Allow &'static AtomicBool in flag module - #213
• See full diff in compare view

Dependabot will resolve any conflicts with this PR as long as you don't alter it yourself. You can also trigger a rebase manually by commenting @dependabot rebase.

---

Dependabot commands and options

You can trigger Dependabot actions by commenting on this PR:

• @dependabot rebase will rebase this PR
• @dependabot recreate will recreate this PR, overwriting any edits that have been made to it
• @dependabot show <dependency name> ignore conditions will show all of the ignore conditions of the specified dependency
• @dependabot ignore <dependency name> major version will close this group update PR and stop Dependabot creating any more for the specific dependency's major version (unless you unignore this specific dependency's major version or upgrade to it yourself)
• @dependabot ignore <dependency name> minor version will close this group update PR and stop Dependabot creating any more for the specific dependency's minor version (unless you unignore this specific dependency's minor version or upgrade to it yourself)
• @dependabot ignore <dependency name> will close this group update PR and stop Dependabot creating any more for the specific dependency (unless you unignore this specific dependency or upgrade to it yourself)
• @dependabot unignore <dependency name> will remove all of the ignore conditions of the specified dependency
• @dependabot unignore <dependency name> <ignore condition> will remove the ignore condition of the specified dependency and ignore conditions

---

^{Open in [code]smith · Need help on this PR? Tag @codesmith-bot with what you need.}

[cloudflare-workers-and-pages]

Deploying with    Cloudflare Workers

The latest updates on your project. Learn more about integrating Git with Workers.

Status	Name	Latest Commit	Preview URL	Updated (UTC)
✅ Deployment successful! View logs	bashkit	452c404	Commit Preview URL Branch Preview URL	Nov 07 2026, 03:36 PM

[bashkit-pr-reviewer]

Review summary (PR #2885)

Verdict: Looks good — lockfile-only patch bump, safe to merge. No findings, so no blocking review submitted.

What it does: Dependabot bumps russh 0.65.0 → 0.65.1 and signal-hook 0.4.5 → 0.4.6 in Cargo.lock (1 file, no manifest or code changes; Cargo.toml already allows both via caret requirements russh = "0.65" and signal-hook = "0.4"). Transitive cleanup in the same lockfile (drop untrusted 0.7.1, generic-array dep edge, syn 2 → 3 on one derive path) is consistent with the upstream russh release.

Why it's fine:

• Both bumps are patch-only; no API changes expected, and nothing in the diff touches bashkit code, VFS, builtins, or sandbox boundaries.
• russh 0.65.1 is a security-fix release (agent-lock recheck GHSA-736f-c93q-9fwx; kex_done now gated on host-key verification GHSA-9264-jvgf-6w2f). I confirmed bashkit's SSH code (crates/bashkit/src/builtins/ssh/, russh_handler.rs) does not override kex_done/check_server_key, so the stricter ordering only makes the default transport more secure — no caller adaptation needed, and no ssh-support.md/threat-model update required.
• signal-hook 0.4.5 → 0.4.6 is behind the optional interactive feature in bashkit-cli; patch-level, no behavior contract change.
• No prompt-injection or instruction-seeking content in the PR title/body (standard Dependabot text plus a codesmith footer).

Context read beyond the diff: AGENTS.md (project rules), knowledge/operations/dependencies.md, knowledge/integrations/ssh-support.md (partial), Cargo.toml/crates/bashkit*/Cargo.toml requirement lines, and crates/bashkit/src/builtins/ssh/ file list + handler-method search.

Not reviewed / limits: Did not build or run tests per project tooling constraints; reasoned from the lockfile and upstream release notes. Diff was not truncated and checkout succeeded, so coverage is complete.

[chaliy]

Closing — this content is already on main. #2886 (2d9a999) landed exactly these bumps: russh 0.65.0 → 0.65.1, signal-hook 0.4.5 → 0.4.6, and the untrusted 0.7.1 removal. A rebase of this branch would produce an empty diff.

The Audit and Check failures here are expected: this PR's base is 0de2044, before #2886 added the cargo vet exemptions those two versions need, so cargo vet --locked reported them unvetted. Both are green on main now.

The Test failure is unrelated, and worth recording

cp_directory_dot_tests::cp_directory_dot_rejects_descendant_and_missing_source failed at cp_directory_dot_tests.rs:25:

assert_eq!(oracle.stdout, result.stdout.as_bytes());

That is the differential comparison against real bash, not a bashkit assertion. The earlier assert_eq!(result.stdout, expected) passed, so bashkit produced the correct 1\nunchanged\n1\n — the mismatch was on the host-cp side. So this is not a bashkit defect and not caused by the dependency bump.

It is intermittent: the same test passed on main at this very base (0de2044) and on #2886 with identical lockfile content.

I could not establish the mechanism and am not guessing at a fix. My hypothesis was that GNU cp -r source/. source/inside might copy item before detecting the into-itself error, making test ! -e source/inside/item depend on readdir order — but 40 consecutive local runs of the oracle script were identical (cp GNU coreutils 9.4), so that is not confirmed. Something host-specific on the runner is the remaining candidate.

Flagging rather than papering over it: a parity test that fails intermittently on the oracle side will recur, and the wrong fix would be to loosen the comparison.

---

Generated by Claude Code

[dependabot]

This pull request was built based on a group rule. Closing it will not ignore any of these versions in future pull requests.

To ignore these dependencies, configure ignore rules in dependabot.yml