fix: resolve review #004 findings W1-W4 + close review gate

W1 (call/protocol/abort-cascade-wiring): wire AbortCascade into
CallAdapter handle_stream for EVENT_ABORTED. Cascades with
AbortPolicy::AbortDependents, aborts root, no descendant frames on
wire (ADR-016 Decision 2). Two integration tests added.

W2 (core/endpoint-client-fingerprint): extract TLS client cert
fingerprint in dispatch_quinn (SHA256:<hex> of leaf cert DER via
peer_identity) and dispatch_iroh (ed25519:<hex> of peer NodeId).
Fingerprint format documented in auth.md. Server config change
(with_no_client_auth → request-but-don't-require) deferred to new
follow-up task core/endpoint-request-client-cert.

W3 (vault/mnemonic-debug-redaction): replace Mnemonic derive(Debug)
with manual redacting impl (phrase: "[REDACTED]"). Seed confirmed
no Debug impl. Redaction test added.

W4 (core/auth-apikey-resources): Option B — drop entry.resources from
spec. External identities (token/fingerprint) grant scopes only;
resource-scoped ACLs are composition-internal (ADR-015/022). auth.md
corrected + limitation documented. Two tests confirm empty resources.

review-post-impl-fixes: all 4 verified, workspace green (326 tests,
0 failures, 0 clippy warnings). Review #004 status → resolved.

Graph: 34 tasks, 12 gens.

tasks/call/protocol/abort-cascade-wiring.md

@@ -1,7 +1,7 @@
 ---
 id: call/protocol/abort-cascade-wiring
 name: Wire AbortCascade into CallAdapter inbound event path (ADR-016)
-status: pending
+status: completed
 depends_on: [call/protocol/abort-cascade]
 scope: narrow
 risk: medium
@@ -127,4 +127,14 @@ frame actually reaches `cascade_abort`.
 > criteria likewise omitted "inbound `call.aborted` triggers cascade."
 > This task closes that integration gap — all the hard logic already
 > exists and is tested; this task adds the ~30-line bolt and the one
-> integration test that would have caught the gap.
+> integration test that would have caught the gap.
+
+## Summary
+
+`handle_stream` now matches `EVENT_ABORTED` → invokes
+`AbortCascade::cascade_abort` with `AbortPolicy::AbortDependents`, then
+aborts the root. Non-requested/non-aborted events still log at `debug!`.
+No descendant `call.aborted` frames sent on the wire. Two integration
+tests: cascade removes parent + child from `PendingRequestMap`; unknown
+request_id is a no-op. `cargo test -p alknet-call` (161 tests) and
+clippy clean.

tasks/core/auth-apikey-resources.md

@@ -1,7 +1,7 @@
 ---
 id: core/auth-apikey-resources
 name: Reconcile ApiKeyEntry.resources — add field to type and populate in resolve_api_key, or drop from spec
-status: pending
+status: completed
 depends_on: []
 scope: narrow
 risk: low
@@ -114,4 +114,17 @@ applied to handler-internal composition identities
 > decision first, then implement. If the decision is A and the
 > implementation is more than ~30 lines, split a follow-up
 > `level: implementation` task (`core/auth-apikey-resources-impl`)
-> depending on this one.
+> depending on this one.
+
+## Summary
+
+Decision: **Option B** — dropped `entry.resources` from the spec.
+Rationale: `Identity.resources` is populated only by
+`CompositionAuthority::as_identity` (the composition path, ADR-015/022).
+All architecture examples use scope-based ACLs for external identities
+(`fs:read`, `vastai:query`, `llm:call`). Adding a second
+resource-population path for API keys would muddy the external/internal
+separation without a demonstrated downstream need. `auth.md:153`
+corrected to `resources: {}`; documented limitation added. Two tests
+confirm both `resolve_api_key` and `resolve_identity_from_fingerprint`
+return empty resources. `cargo test -p alknet-core` and clippy clean.

tasks/core/endpoint-client-fingerprint.md

@@ -1,7 +1,7 @@
 ---
 id: core/endpoint-client-fingerprint
 name: Extract TLS client certificate fingerprint in endpoint dispatch (ADR-004)
-status: pending
+status: completed
 depends_on: []
 scope: narrow
 risk: medium
@@ -115,4 +115,17 @@ in the codebase before committing.
 > client certs. The extraction code is correct either way — it returns
 > `None` when no cert was presented, which is the current behavior, so
 > landing extraction first is a safe no-op until the server config
-> changes.
+> changes.
+
+## Summary
+
+Added `extract_quinn_client_fingerprint` (leaf cert DER → `SHA256:<hex>`
+via `peer_identity()` downcast) and `extract_iroh_client_fingerprint`
+(peer `NodeId` → `ed25519:<hex>`). Both dispatch functions now pass the
+extracted fingerprint to `build_auth_context`. Fingerprint format
+documented in `auth.md` (table: quinn X.509 vs iroh raw Ed25519).
+Server config still uses `with_no_client_auth()` — extraction is a safe
+no-op. Follow-up task `core/endpoint-request-client-cert` created for the
+server config change. Two unit tests cover fingerprint format +
+determinism. `cargo test -p alknet-core --all-features` (59 tests) and
+clippy clean.

tasks/core/endpoint-request-client-cert.md

@@ -0,0 +1,90 @@
+---
+id: core/endpoint-request-client-cert
+name: Switch rustls ServerConfig from with_no_client_auth to request-but-don't-require client certs
+status: pending
+depends_on: [core/endpoint-client-fingerprint]
+scope: narrow
+risk: medium
+impact: component
+level: implementation
+---
+
+## Description
+
+`core/endpoint-client-fingerprint` landed the extraction logic: when a
+client certificate *is* presented, `dispatch_quinn` / `dispatch_iroh`
+extract the fingerprint and populate `AuthContext`. However, the server
+still builds `rustls::ServerConfig` with `with_no_client_auth()` in all
+three `TlsIdentity` branches (`endpoint.rs:477`, `490`, `501`), so the
+server never *requests* a client cert. Extraction is a safe no-op until
+this task changes the server-side TLS config.
+
+This follow-up switches from `with_no_client_auth()` to a
+request-but-don't-require mode so that peers presenting a client cert
+(X.509 or RFC 7250 raw Ed25519 key) flow through the extraction path
+landed in the predecessor task, while peers without a cert still connect
+without regression.
+
+### Design decision: how to request-but-not-require
+
+rustls does not have a direct `with_optional_client_auth()` builder.
+The standard approach is:
+
+1. Build the config with `.with_client_auth(verifier)` where `verifier`
+   is a custom `ServerCertVerifier` that accepts any presentation (returns
+   `Ok(Certified::yes())` when a cert is presented, `Ok(Certified::no())`
+   when none is presented — rustls 0.23's `WebPkiServerVerifier` cannot
+   be used directly for optional auth).
+2. Alternatively, use `rustls::server::WebPkiServerVerifier` with a
+   `NoClientAuth` fallback — check the exact rustls API available in the
+   pinned version before implementing.
+
+Read the rustls API docs for the pinned version
+(`rustls::server::ServerConfig::builder_with_provider`) and confirm the
+correct verifier construction. The key property: a peer *may* present a
+cert, and if it does, `peer_identity()` returns it; if it doesn't, the
+connection still succeeds.
+
+### iroh path
+
+iroh's `Endpoint` builder uses its own TLS session internally. For the
+raw-key path (`TlsIdentity::RawKey`), iroh already advertises
+`only_raw_public_keys()` via `RawKeyCertResolver` — the server-side half
+of RFC 7250. The client-side presentation is set by the client's
+`rustls::ClientConfig`, not the server. So the iroh path may already
+receive peer identities when the client is an iroh node (the `NodeId` is
+always in the TLS cert). Verify: does `Connection::remote_node_id()`
+already work for iroh connections today, or does it require the server to
+request client certs? If iroh always presents a cert (raw-key mode), no
+server-side change is needed for the iroh path — only quinn/X.509 needs
+this task. Confirm before implementing.
+
+## Acceptance Criteria
+
+- [ ] `build_rustls_server_config` uses request-but-don't-require client auth (not `with_no_client_auth()`) for at least the X.509 path
+- [ ] Peer presenting a client cert: `peer_identity()` returns the cert chain → fingerprint extraction works end-to-end
+- [ ] Peer without a client cert: connection still succeeds, `tls_client_fingerprint` is `None` (no regression)
+- [ ] iroh path: confirm whether a server-side change is needed; if yes, apply it; if no, document why
+- [ ] Integration test: quinn endpoint with a client that presents a cert → `AuthContext.tls_client_fingerprint` is `Some(SHA256:...)`
+- [ ] Integration test: quinn endpoint with a client that presents no cert → `AuthContext.tls_client_fingerprint` is `None` and connection succeeds
+- [ ] `cargo test -p alknet-core --all-features` succeeds
+- [ ] `cargo clippy -p alknet-core --all-features --all-targets` succeeds with no warnings
+- [ ] `auth.md` updated: server-config decision documented (request-but-don't-require, not no-client-auth)
+
+## References
+
+- tasks/core/endpoint-client-fingerprint.md — predecessor task (landed extraction, deferred this config change)
+- crates/alknet-core/src/endpoint.rs:477, 490, 501 — the three `with_no_client_auth()` sites
+- crates/alknet-core/src/endpoint.rs — `extract_quinn_client_fingerprint` / `extract_iroh_client_fingerprint` (already landed, waiting for certs to flow)
+- docs/architecture/crates/core/auth.md — fingerprint format table and endpoint-level resolution flow
+- docs/architecture/decisions/004-auth-as-shared-core.md — ADR-004 (hybrid resolution)
+- docs/architecture/open-questions.md — OQ-12 (TLS identity provisioning)
+
+## Notes
+
+> Split from `core/endpoint-client-fingerprint` per the task's own
+> suggestion: extraction is correct either way (returns `None` when no
+> cert), so landing it first is a safe no-op. This task is the
+> behavioral change that makes fingerprints actually flow. The risk is
+> medium because it alters the TLS handshake for every connection —
+> ensure the no-cert-peer case has explicit test coverage.

tasks/review-post-impl-fixes.md

@@ -1,7 +1,7 @@
 ---
 id: review-post-impl-fixes
 name: Review the four post-implementation sanity-check #004 fixes for spec conformance
-status: pending
+status: completed
 depends_on: [call/protocol/abort-cascade-wiring, core/endpoint-client-fingerprint, vault/mnemonic-debug-redaction, core/auth-apikey-resources]
 scope: moderate
 risk: low
@@ -93,4 +93,23 @@ check.md`, does not introduce new spec drift, and is adequately tested.
 > at the end of a fix batch, with `scope: moderate`, `risk: low`,
 > `impact: phase`. It does not need to re-derive the findings — review
 > #004 already did that work. It only needs to confirm the fixes land
-> correctly and the workspace stays green.
+> correctly and the workspace stays green.
+
+## Summary
+
+All four fixes verified against acceptance criteria:
+- W1: `handle_stream` handles `EVENT_ABORTED`, cascades with
+  `AbortDependents`, no descendant frames on wire, root removed, two
+  integration tests pass.
+- W2: both dispatch paths extract fingerprints, format documented in
+  `auth.md`, no-cert case returns `None` (no regression), server-config
+  change deferred to `core/endpoint-request-client-cert`.
+- W3: `Mnemonic` has manual redacting `Debug`, `Seed` has no `Debug`,
+  redaction test passes.
+- W4: Option B — spec corrected, limitation documented, both resolvers
+  return empty resources, tests pass.
+
+Workspace green: `cargo build --workspace --all-features` ✓, `cargo test
+--workspace --all-features` (326 tests, 0 failures) ✓, `cargo clippy
+--workspace --all-features --all-targets` (0 warnings) ✓. Review #004
+status updated to `resolved`.

tasks/vault/mnemonic-debug-redaction.md

@@ -1,7 +1,7 @@
 ---
 id: vault/mnemonic-debug-redaction
 name: Replace Mnemonic derive(Debug) with redacting impl to prevent seed phrase leak
-status: pending
+status: completed
 depends_on: []
 scope: single
 risk: low
@@ -107,4 +107,12 @@ fn test_mnemonic_debug_redacts_phrase() {
 
 > Small fix, but eliminates a latent root-of-trust leak. The same
 > pattern (custom redacting `Debug`) is already established in three
-> other places in this codebase — this task brings `Mnemonic` in line.
+> other places in this codebase — this task brings `Mnemonic` in line.
+
+## Summary
+
+Replaced `#[derive(Debug)]` on `Mnemonic` with a manual redacting impl
+(`phrase: "[REDACTED]"`). Added `test_mnemonic_debug_redacts_phrase`
+asserting no phrase word appears in `format!("{:?}", mnemonic)`.
+Confirmed `Seed` has no `Debug` impl (derives only `Clone, Zeroize`).
+`cargo test -p alknet-vault` and clippy clean.