Sign the runtime's own gate proposals with the operator identity (SPEC-006) - #388
Merged
Conversation
…C-006) Closes the last "gate record unsigned" gap. The self-extend bridge recorded its automatic proposals through a signer-less GateStore, so a cycle's verdict carried no provenance while an operator's manual decision (via `ashlar gates`) did. Now the bridge loads the operator identity and signs, so both halves of admission — the runtime's proposals and the human's decisions — carry the same signature when a key is present, and stay unsigned when none is (S-2). - TryRecordAsync gains an optional `signer` seam: production loads the machine key (OperatorKey.TryLoad), tests inject an identity so they never depend on ~/.ashlar/keys. - A corrupt operator key fails the record LOUD — the same "GATE ERROR" shape as an unreadable policy — rather than silently recording unsigned, which would hide that signing was expected. - The production caller (SelfExtendRunnerAdapter) is unchanged: the new param is optional and trailing, so it defaults to the machine key. Ashlar.Tests.BackgroundAgents 513/513 on net8.0 and net10.0, including a new test that a runtime proposal is signed (read back by a keyless store, which verifies via the embedded key). Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for freeto join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Closes the last "gate record unsigned" gap in SPEC-006. The self-extend bridge recorded its automatic proposals through a signer-less
GateStore, so a cycle's verdict carried no provenance while an operator's manual decision (viaashlar gates, #387) did. Now the bridge loads the operator identity and signs, so both halves of admission — the runtime's proposals and the human's decisions — carry the same signature when a key is present, and stay unsigned when none is (S-2).What changed (src/ only — kernel-first, clean layer gate)
SelfExtendAdmissionBridge.TryRecordAsyncopens the store withOperatorKey.TryLoad()and gains an optionalsignerseam: production loads the machine key, tests inject an identity so they never depend on~/.ashlar/keys.GATE ERRORshape as an unreadable policy — never silently unsigned.SelfExtendRunnerAdapter) is unchanged: the new param is optional and trailing.Verification
Ashlar.Tests.BackgroundAgents513/513 on net8.0 and net10.0, including a new test that a runtime proposal is signed and reads back (a keyless reader still verifies via the record's embedded key). No regressions in the existing bridge/M1 enforcement tests.Next
With this, every gate record is signed when a key is present. The remaining SPEC-006 surface is user-visible:
ashlar verifystill printsunsigned— making it show a real signer fingerprint is the SPEC-003 ledger slice.