chore(rust): upgrade ed25519-dalek 2.x -> 3.x and rand 0.8 -> 0.10 (fixes #3355)

[AlgoVoi (Christopher Hopley) (chopmob-cloud)]

Summary

Upgrades ed25519-dalek 2.2.0 -> 3.0.0 and rand 0.8.6 -> 0.10.2 across the agent-governance-rust workspace, as requested in #3355.

Dependabot #3269 (dalek 3) and #3271 (rand 0.10) each failed build-rust because they bump one crate at a time. The two must move together: ed25519-dalek 3 rides rand_core 0.9, which rand 0.10 provides, so bumping either alone leaves a rand_core version mismatch. This is a single coordinated PR.

What actually changed

Bumping both together showed that the ed25519-dalek 2 -> 3 signing/verifying surface used in this crate is source-compatible (SigningKey::generate, Signature::from_bytes, VerifyingKey::from_bytes, Signer/Verifier). All the real edits are the rand 0.9/0.10 trait and module reshuffle:

• rand::distributions -> rand::distr (clock.rs)
• rand::thread_rng() -> rand::rng() in non-key-material paths (clock.rs, the AES-GCM nonce in credential_vault.rs)
• the old Rng extension trait (providing sample_iter) is now RngExt (clock.rs)
• the old RngCore core trait (providing fill_bytes) is now named Rng (credential_vault.rs)
• rand::rngs::OsRng was removed; key generation now uses UnwrapErr(SysRng) (identity.rs, identity_support.rs, credential_vault.rs), see below

Note on key-generation RNG

SigningKey::generate in ed25519-dalek 3 requires an infallible CryptoRng (pub fn generate<R: CryptoRng + ?Sized>(csprng: &mut R)). The documented pattern in the ed25519-dalek 3.0.0 SigningKey::generate docs is OS entropy through the UnwrapErr adapter: let mut csprng = UnwrapErr(SysRng); SigningKey::generate(&mut csprng). SysRng (the getrandom 0.4 system source re-exported by rand 0.10 as rand::rngs::SysRng) implements the fallible TryCryptoRng, and rand_core's UnwrapErr wrapper turns it into an infallible CryptoRng via the blanket impl, satisfying the bound directly from OS entropy with no new dependency and no lockfile change.

Every key-material path draws from UnwrapErr(SysRng): AgentIdentity::generate/delegate, Credential::issue, KeyRotationManager::rotate, and the vault's AES-256-GCM CredentialVault::generate_key. Thread RNG (rand::rng()) remains only where key material is not involved (the AES-GCM nonce, generated identifiers, the attestation challenge nonce, and the MCP clock nonce). This draws long-lived key material straight from the OS CSPRNG and removes the thread-local reseed and fork-safety caveats from the keygen path. The dependency-audit doc records the split.

Validation

agent-governance-rust, from a green baseline:

	baseline (2.2.0 / 0.8.6)	this PR (3.0.0 / 0.10.2)
cargo build --workspace	clean	clean, no new warnings
cargo test --workspace --locked	514 pass, 0 fail	518 pass, 0 fail (Linux)
cargo clippy --workspace --all-targets	9 warnings	9 warnings (0 new)

The PR adds two keygen tests (distinct usable keys with cross-verification rejection; distinct non-zero vault keys), which is the +4 delta over the 514 baseline together with the branch's earlier additions. The suite still includes the signature-rejection tests (trust::test_verify_peer_rejects_mismatched_claimed_peer, trust::test_verify_peer_rejects_signature_not_created_by_peer, mcp::signing::tests::rejects_replayed_messages), so verification still refuses forged and replayed signatures, not merely accepts valid ones.

Fixes #3355.

[azure-pipelines]

Azure Pipelines:
There may be pipelines that require an authorized user to comment /azp run to run.

[github-actions]

🤖 AI Agent: code-reviewer — View details

AI-generated review output. Treat it as untrusted analysis and verify before acting.

TL;DR: 0 blockers, 0 warnings. Safe and clean upgrade.

#	Sev	Issue	Where

No issues found. Clean change.

[github-actions]

🤖 AI Agent: contributor-guide — View details

AI-generated review output. Treat it as untrusted analysis and verify before acting.

Welcome, and thank you for your contribution! Great job coordinating the ed25519-dalek and rand upgrades to resolve dependency conflicts effectively.

Before we can merge, please address the following:

1. Ensure that all removed dependencies in Cargo.lock are no longer required by any part of the project.
2. Verify that the changes to rand usage (e.g., rand::rng() and RngExt) are consistent across all affected files.

For more details, refer to our CONTRIBUTING.md. Let us know if you need any help!

[github-actions]

🤖 AI Agent: docs-sync-checker — Docs Sync

AI-generated review output. Treat it as untrusted analysis and verify before acting.

Docs Sync

• README.md -- no updates found for the changes in the pull request.
• CHANGELOG.md -- missing entry for the upgrade of ed25519-dalek and rand versions, as well as the associated behavioral changes (e.g., OsRng to rand::rng() for key generation).

Please ensure the CHANGELOG.md reflects these updates.

[github-actions]

🤖 AI Agent: security-scanner — View details

AI-generated review output. Treat it as untrusted analysis and verify before acting.

No security issues found.

[github-actions]

🤖 AI Agent: test-generator — View details

AI-generated review output. Treat it as untrusted analysis and verify before acting.

Test coverage looks good. No gaps identified.

[github-actions]

🤖 AI Agent: breaking-change-detector — API Compatibility

AI-generated review output. Treat it as untrusted analysis and verify before acting.

API Compatibility

Severity	Change	Impact
High	Upgrade ed25519-dalek from 2.x to 3.x	Potential breaking change due to major version bump. While the PR notes that the signing/verifying surface is source-compatible, downstream consumers relying on internal or undocumented behavior may face issues.
High	Upgrade rand from 0.8 to 0.10	Potential breaking change due to major version bump. Changes in trait and module structure (e.g., rand::distributions → rand::distr, OsRng replaced with ThreadRng) may impact consumers relying on the previous API.
Medium	Replacement of rand::rngs::OsRng with rand::rng()	May affect users who explicitly depend on OsRng for cryptographic random number generation.

[github-actions]

PR Review Summary

Check	Status	Details
🔍 Code Review	⚠️ Missing	No current-run comment
🛡️ Security Scan	⚠️ Missing	No current-run comment
🔄 Breaking Changes	⚠️ Missing	No current-run comment
📝 Docs Sync	⚠️ Missing	No current-run comment
🧪 Test Coverage	⚠️ Missing	No current-run comment

Verdict: ⚠️ AI review incomplete; ready for human review

AI review comments are untrusted advisory output. The summary reports workflow-generated completion status only, not model-authored pass/fail claims.

[github-actions]

🔴 Contributor Check: HIGH

Check	Result
Profile	HIGH
Credential	LOW
Overall	HIGH

Automated check by AGT Contributor Check.

[Copilot]

Pull request overview

TL;DR: 0 blockers, 0 warnings. No issues found. Clean change.

Changes:

• Upgrade ed25519-dalek from 2.2.0 to 3.0.0 and rand from 0.8.6 to 0.10.2 across the agent-governance-rust workspace.
• Update call sites for rand 0.10 API moves/renames (e.g., distributions → distr, thread_rng() → rng(), Rng → RngExt, RngCore → Rng).

Reviewed changes

Copilot reviewed 5 out of 6 changed files in this pull request and generated no comments.

Show a summary per file

File	Description
agent-governance-rust/Cargo.toml	Pins ed25519-dalek to =3.0.0 and rand to =0.10.2 at the workspace level.
agent-governance-rust/Cargo.lock	Updates the resolved dependency graph for the coordinated ed25519-dalek/rand upgrade.
agent-governance-rust/agentmesh/src/identity.rs	Switches key generation RNG call from OsRng to rand::rng() for SigningKey::generate.
agent-governance-rust/agentmesh/src/identity_support.rs	Updates SigningKey::generate RNG call sites to rand::rng() for credential/key rotation paths.
agent-governance-rust/agentmesh/src/credential_vault.rs	Adapts RNG trait usage (RngCore → Rng) and updates thread_rng() → rng() for byte filling.
agent-governance-rust/agentmesh-mcp/src/mcp/clock.rs	Updates rand imports (distr, RngExt) and RNG creation (rng()) for nonce generation.

[Imran Siddique (imran-siddique)]

MohammadHaroonAbuomar this is the closest Rust PR to merge: mergeable, CI clean, resolves #3355, and the duplicate #3418 is now closed. Could you give it a code-owner review when you have a moment? Thanks.

[Copilot]

Pull request overview

Copilot reviewed 6 out of 7 changed files in this pull request and generated no new comments.

[Copilot]

Pull request overview

Copilot reviewed 7 out of 8 changed files in this pull request and generated no new comments.

[MohammadHaroonAbuomar]

• Process: real CI has never run on this head (all green checks are pull_request_target bot jobs; build-rust/CodeQL/SBOM stuck action_required) and the PR is CONFLICTING on the exact Cargo.toml/lock block being upgraded. Rebase onto main, get workflow runs approved, require genuinely green CI. (Local compensating run at head: cargo test 514/514 pass; code content verified sound incl. dalek3 signing determinism and no seeded-RNG leaks.)

Minor:

• audit doc omission: OsRng->ThreadRng is not strictly equivalent (thread-local ChaCha12, reseeds per 64KiB, not fork-safe; no fork usage in workspace today). Add a line.

[MohammadHaroonAbuomar]

• Process: real CI has never run on this head (all green checks are pull_request_target bot jobs; build-rust/CodeQL/SBOM stuck action_required) and the PR is CONFLICTING on the exact Cargo.toml/lock block being upgraded. Rebase onto main, get workflow runs approved, require genuinely green CI. (Local compensating run at head: cargo test 514/514 pass; code content verified sound incl. dalek3 signing determinism and no seeded-RNG leaks.)

Minor:

• audit doc omission: OsRng->ThreadRng is not strictly equivalent (thread-local ChaCha12, reseeds per 64KiB, not fork-safe; no fork usage in workspace today). Add a line.

[MohammadHaroonAbuomar]

• .cspell.json:105: the rebase merge dropped "starlette" relative to main (added there by #3411/#3085). The word is used in agent-mesh (errors.py, http_middleware.py, request_auth.py, CHANGELOG.md), so the next PR touching those lines would fail Spell-check. Please re-add "starlette", to the words list.

[MohammadHaroonAbuomar]

Correction on the .cspell.json item above: main lists starlette twice (once earlier in the list, once near line 105), and this PR drops only the duplicate, so nothing is lost. Please ignore that ask; the audit-table refresh is the only item left.

[AlgoVoi (Christopher Hopley) (chopmob-cloud)]

Thanks for the careful read. I think starlette is actually still covered: it remains in the top-level words array of .cspell.json (line 69 on this branch), so cspell still accepts it and the Spell-check job is green on the current head. What the diff shows at line ~102 is a position move rather than a removal: the rebase dropped the entry at main's line and the branch keeps its own earlier one, so there is exactly one starlette in the list either way, and the agent-mesh usages (errors.py, http_middleware.py, request_auth.py, CHANGELOG.md) stay covered after merge.

Happy to relocate it to match main's ordering if you would prefer the list stay positionally aligned to avoid future diff noise, just let me know.

[Ricky Gummadi (Ricky-G)]

@AlgoVoi can you please rebase, resolve conflicts and push for review again, if its still stale for another week, this will have to come back in as a new PR.

--PR-older-than-a-month

[MohammadHaroonAbuomar]

AlgoVoi (Christopher Hopley) (@chopmob-cloud), a status note so you have the full picture. The one open item on this PR is the audit table at docs/dependency-audits/2026-07-27-ed25519-dalek-3-rand-0.10.md:38, plus a rebase (the branch now conflicts with main). Meanwhile #4104 by a maintainer implements the same upgrade; two of its files are byte-identical to yours and it now credits you as co-author. The earlier ping on this thread went to @AlgoVoi, which is not your GitHub login, so you may not have seen it.

If you can rebase and fix the audit table in the next few days, this PR is the one that should land, and the OS-entropy hardening from #4104 can follow on top. If you would rather not, say so and #4104 goes in with your credit. Either way, thank you for carrying this since July.

[MohammadHaroonAbuomar]

• agent-governance-rust/Cargo.lock:1 The rebase carries a downgrade: main has cedar-policy, cedar-policy-core and cedar-policy-formatter at 4.13.0, and this lock resolves them at 4.12.0, so a squash merge would move main backwards on those three crates. Please regenerate the lock against current main (for example cargo update -p cedar-policy -p cedar-policy-core -p cedar-policy-formatter --precise 4.13.0, or re-resolve from main's lock and re-apply only the dalek and rand changes) so the diff touches only the crates this upgrade changes.

[AlgoVoi (Christopher Hopley) (chopmob-cloud)]

MohammadHaroonAbuomar Done, this is rebased onto current main and mergeable again (head 51540a67).

• Rebase, no downgrades. The lockfile was regenerated from main, so the diff touches only the dalek/rand upgrade tree. cedar-policy/-core/-formatter stay at 4.13.0 and opentelemetry at 0.33.0 (the stale branch would have moved both backwards), and no unrelated crate changes. chacha20 resolves to 0.10.2, so the yanked 0.10.1 is avoided.
• Audit table refreshed to match the regenerated lock: getrandom, zerocopy/zerocopy-derive, wasi, digest and sha2 are unchanged by this upgrade (main has moved since the doc was first written, so the earlier "removed"/"gains a major" rows were stale), and const-oid collapses 0.9.6 + 0.10.2 to 0.10.2.
• Local verification at the rebased head: cargo build, cargo test and cargo clippy are green, full workspace suite 578 passed / 0 failed. The fork's build-rust/CodeQL/SBOM jobs are still action_required pending a maintainer approving the workflow run.

Happy to have the OS-entropy hardening from #4104 follow on top once this lands, per your note. Thanks for carrying the review.

[MohammadHaroonAbuomar]

Approving at 51540a6. The lock now moves nothing backwards: ed25519-dalek, ed25519, curve25519-dalek, signature, fiat-crypto and rand go up, two duplicate crates collapse to their newer version, chacha20 0.10.2 is new (0.10.1 is yanked), six transitive crates drop out, and cedar-policy and opentelemetry stay at main's versions. Every newly pinned version is past the seven-day rule. The audit document's table and prose match the lock row for row. Key generation at all five sites uses SigningKey::generate(&mut UnwrapErr(SysRng)) or SysRng.fill_bytes, the stateless OS source the dalek 3 documentation prescribes, and it panics rather than degrading if the OS source fails; signing and verification paths are unchanged and the rejection tests still pass; the new tests cover distinct usable keys and cross-verify rejection. 578 workspace tests pass locally, the vendored-patch audit passes, and all 18 checks are green. Twelve signed commits. Thank you for carrying this from #3355 through two rebases.

For the record: cargo fmt --check and clippy -D warnings fail on main in files this PR does not touch, and CI does not gate them; that is a separate cleanup. #4104 carries the same upgrade on an older base and should close as superseded once this lands.