fix(attest): correct single-upstream fingerprint keying before tagging v0.43.0#169
Conversation
… non-P256 EC keys
A named single upstream (--upstream NAME=CMD) collapses into the "default"
slot inside VaaraMCPProxy, and _REQUEST_UPSTREAM resolves to "default" at
runtime. The attestation emitter was keyed on the operator-supplied name,
so fingerprint_for("default") missed the precomputed cmd-hash and emitted a
cmd:sha256:unknown-default placeholder for the documented NAME=CMD single-
upstream form. _attest_upstreams_for_slots keys the emitter the same way the
proxy slots upstreams; multi-upstream fan-out keeps operator names.
Also reject EC signing keys that are not P-256: ES256 emits a fixed 32-byte
r||s, so a non-secp256r1 curve would be mislabeled ES256 and then silently
fail to sign. Fail up front with a clear AttestConfigError.
Adds 5 tests: slot-keying (named-single collapses, multi preserved), curve
validation (P-384 rejected, P-256 accepted), and end-to-end fingerprint.
The marketplace ref had been stale at v0.41.0 across v0.42.0 and v0.43.0, so plugin install pinned users two releases back. Point it at the release being tagged.
|
Warning Review limit reached
More reviews will be available in 6 minutes and 21 seconds. Learn how PR review limits work. Your organization has run out of usage credits. Purchase more in the billing tab. ⌛ How to resolve this issue?After more reviews become available, a review can be triggered using the We recommend that you space out your commits to avoid hitting the rate limit. 🚦 How do rate limits work?CodeRabbit enforces hourly rate limits for each developer per organization. Our paid plans include higher PR review limits than trial, open-source, and free plans. In all cases, reviews become available again over time. During sustained high-volume PR review activity, CodeRabbit may temporarily slow when the next review becomes available. Please see our Fair Usage Limits Policy for further information. ℹ️ Review info⚙️ Run configurationConfiguration used: defaults Review profile: CHILL Plan: Pro Plus Run ID: 📒 Files selected for processing (5)
✨ Finishing Touches🧪 Generate unit tests (beta)
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
1 participant
What this is
v0.43.0 merged to
mainin #168 but was never tagged, so PyPI, the GH Release, and the registry slots never fired. Before finishing the tag I found two correctness bugs in the merged attestation code, plus a stale plugin marketplace ref. This lands the fixes so the published v0.43.0 is correct from the first tag, rather than shipping a broken artifact and chasing it with a v0.43.1.Fixes
Single-upstream fingerprint keying. A named single upstream (
--upstream NAME=CMD) collapses into the"default"slot insideVaaraMCPProxy, and_REQUEST_UPSTREAMresolves to"default"at runtime. The attestation emitter was keyed on the operator-supplied name, sofingerprint_for("default")missed the precomputed command hash and emitted acmd:sha256:unknown-defaultplaceholder for the documentedNAME=CMDform._attest_upstreams_for_slotsnow keys the emitter the way the proxy slots upstreams. Multi-upstream fan-out keeps the operator names.Reject non-P-256 EC signing keys. ES256 emits a fixed 32-byte
r||s, so a non-secp256r1 curve would be mislabeledES256and then silently fail to sign (the emit path swallows signing errors). The key loader now rejects it up front with a clearAttestConfigError.Chore
Marketplace ref.
.claude-plugin/marketplace.jsonhad been stale atv0.41.0across v0.42.0 and v0.43.0, pinning plugin installs two releases back. Bumped tov0.43.0.Tests
5 new tests: slot-keying (named-single collapses to default, multi preserved), curve validation (P-384 rejected, P-256 accepted), and an end-to-end fingerprint check. Affected suites: 61 passed, ruff clean.