Preserve known non-skill metadata in projection commitments, reject duplicate and oversized bundles, close Codex exec grammar ambiguity, and keep raw-bundle requirements explicit. Keep layer-2 Sage materialization tests out of this layer.
Validated by the exact layer-1 build and 75-test suite; Apple lifecycle remains a separately skipped layer-2 concern.
Signed-off-by: 0xkite <me@hack4kitesdev.dev>
Mixed libraries were failing start because any non-skill entry aborted the
projection. The managed sandbox bypass also has to sit after `exec` so
headless Codex can run inside Glove.
Model and PTY cannot self-report pass or failure. Glove now runs
each refinement-eval-v1 session through a declarative evaluator that
checks bounded fixture assertions (termination cause, exit code, max
latency, required and forbidden transcript literals) against audited
resource events, then binds the evaluated outcome inside the
authenticated receipt digest. Capability advances to version 1 with
the evaluator landed.
Keep credentials and network access behind authenticated, audited brokers while binding launch adapters, immutable closures, recovery identities, receipts, and cleanup to the approved plan. Add Apple Container and Linux release lanes without weakening the outer sandbox.
Detect and bind operator-installed harnesses through explicit, digest-bound runtime policy. Keep macOS and Linux shipping lanes capability-aware and fail closed when isolation evidence is absent.
