[amad-malik]

Summary

DRAFT C1 source composition at d3a9311ed9032f2fb5b9349b69ea0473446e465b, tree c1e414cfd9c46a4c22860c5ff4b96fefa88045f6. Exact Nix C1-01/C1-02 input 9e11c9939aca18adfedd91ce09447372e0990716 is merged cleanly into 8638e0cb; only its project-key source/test blobs differ from that checkpoint. This head is FROZEN for Kai's existing Hale component-A closure check; no further push during review.

open_unlock
  -> recheck original leaf, ACL, and held-directory predicates after read
  -> issue descriptor only while the original session deadline remains valid
real make_secret -> Core custody -> finite bootstrap -> cleanup/close/record

All ancestry and the committed causal join test are preserved, together with Core/contracts/schema/CI source, existing writer inventory and all21 original migration bytes. No new mechanism, suite registration or unrelated edit.

Evidence

• Before: Nix committed seven assertion-RED C1 controls with0 setup errors on both Pythons. The32-file owner index is hash-verified and the tests are byte-identical between his RED and fix commits. These retained input receipts receive no new-head PASS credit.
• After: EACH local Python3.13.7/3.14.0 passes847 tests (294runtime +38schema +225installer +111renderer +15CLI +103strictness +61inventory). Each225-test installer run includes all nine C1 controls; the111 renderer tests include the unchanged real-resolver causal join. Separate writer audits classify162 writers with0 unclassified/missing/stale each. NONSECRET source fixtures use owned internal0700 temporary homes and are removed.
• Exact-head ci39197852285 SUCCESS6/6; next-offline39197852261 SUCCESS18/18; all24 PR checks SUCCESS. Python 3.13.7: 255 suites, 3848 selected = 3688 PASS + 160 executed approved assertion declarations with ZERO PASS credit; 3869 independently collected, 21 unselected with zero execution credit; Python 3.14.0: 255 suites, 3848 selected = 3688 PASS + 160 executed approved assertion declarations with ZERO PASS credit; 3869 independently collected, 21 unselected with zero execution credit. Independent collections match all actual job partitions and every phase outcome is verified. All nine C1 controls and the unchanged causal join execute and PASS on both hosted Pythons.
• All24 artifact archives/2902 members are hashed and match GitHub server digests. Evidence: /Users/amadmalik/DevVault/helix/output/Cortex/finn-c1-fixes-v7-20261019. Earlier v6 report and425 indexed files remain unchanged.

Original J1/J2 acceptance and the complete finite project-claim/admission packet remain open. J2/008, J3/J4 implementation, actual key/host/targetPG/native/installed effects and signed-helper closure remain held. The module alias in the causal source test supplies dynamic test identity only. Bob's prior S4/S5 findings retain their existing disposition. Kai owns the existing ONE Hale A closure check and canonical merge; this remains a draft.

Merge Danger

Door: two-way for source rollback; persistent effects require separate admission.

Blast Radius: custody

Credential failures

None in this continuation. An early artifact precheck returned workflow-in-progress, downloaded nothing and received no test credit; its raw result and corrected progress statement are retained in METADATA_ATTEMPTS.md.

Open items for Kai

Hale's existing C1-01/C1-02 closure check at this exact frozen head, then Kai's merge decision. No new reviewer, effect admission or J3 release requested. LEAN: reused exact owner fixes, original predicates, causal test and existing qualification; did NOT add a parser/backend/config path/suite; could merge none within this source commission.