Default managed SOCKS hostname DNS to verified HTTPS with literal-IP bootstrap and no plaintext fallback. Supervise the loopback forwarder with private anchored readiness and preserve foreign markers. Include configuration, deployment, documentation and offline checks.
Co-authored-by: GPT-6 <gpt-6@openai.com>
Record supplied fd25cb8 managed-image watch, representative post-warmup media transfers and degraded recovery continuation without claiming avatar or durable family qualification. Require explicit Dockerfile.target and positive entrypoint/marker checks before image promotion; preserve the user README badge.
Co-authored-by: GPT-6 <gpt-6@openai.com>
An isolated SHA-verified replay reproduces exit 78 at managed_experimental_recovery.go:148 with the nested required-hooks guard at managed_transaction.go:69. The production recheck closure retained the adapter path-only transaction instead of the job-bound transaction. Pass the current transaction explicitly without weakening ownership, Lua, runtime or reconciliation guards.
Regression starts recovery with path-only hooks and checks bound ownership for cold/adaptive rechecks. Full native Go tests, affected race tests, go vet and diff checks passed. Corrected image replay and production restoration remain unverified.
Co-authored-by: GPT-6 <gpt-6@openai.com>
Canonicalize validated managed proxy schemes to socks5h for Player, metadata, resource and warmup probes; retain listener ownership, TLS/privacy checks and legacy CLI/pinned checker semantics. Stop canceled inventory requests before Player and metadata fallback, with explicit cancellation and runner-only DNS regressions.
Verification: full native Go tests, affected race tests, go vet and static/config/supervision fixtures passed. An extra whole-package race attempt exceeded its 180s deadline in unrelated TestTuneCLI; affected race coverage completed separately.
Co-authored-by: GPT-6 <gpt-6@openai.com>
Type decoder payload rejections once and share the image nonpass classifier across live health and owned verification. Retain strict qualification, malformed metadata, source-kind separation and mixed safety vetoes; exercise seed fallback, joined workers, fresh evidence and durable absence cooldown.
Co-authored-by: GPT-6 <gpt-6@openai.com>
Keep verified curl peer TLS negatives unqualified without making them process-wide safety failures. Preserve local proxy/configuration, guard, cleanup and mixed cancellation vetoes; test owned verification, rollback and durable absence cooldown.
Co-authored-by: GPT-6 <gpt-6@openai.com>
Fall back only on missing homepage metadata, preserve private bounded inventory and fail-closed error classification. Record supplied field results without claiming image or family qualification.
Co-authored-by: GPT-6 <gpt-6@openai.com>
Preserve operator-owned profiles while rechecking fresh media and bounded native relearning in the deployed daemon. Keep mixed safety failures fatal and expose active coverage separately from qualification. Document source-only acceptance and retain existing recovery budgets and privacy boundaries.
Co-authored-by: GPT-6 <gpt-6@openai.com>
Reuse managedMonitorDatabase to validate the canonical target-owned database and private files/sidecars before openStore can create, chmod or migrate anything. Keep ordinary history read-only and experimental authorization unchanged. Cover unsafe paths, links, modes, owners and sidecars plus valid existing stores.
Co-authored-by: GPT-6 <gpt-6@openai.com>
Use a dedicated writable database dependency only for guarded experimental insertion. Ordinary managed history remains read-only; discovery and authorization gates are unchanged. Reproduced SQLite readonly failure in an isolated copy with private startup logs retained and the restored live target untouched.
Co-authored-by: GPT-6 <gpt-6@openai.com>
Bind ordinary candidate-ID CLI activation to a transaction-local prohibition on experimental policy. Resolve and reject both requested and inherited experimental scope while holding the journal lock, leaving guarded operator activation and trusted monitor/reconciliation behavior unchanged. Cover deterministic operator-policy interleaving without timing-dependent concurrency.
Co-authored-by: GPT-6 <gpt-6@openai.com>
Reject explicit candidate-ID and inherited-policy CLI bypasses of the guarded native-circular operator path. Validate empty private runtime-owned request inodes without following links before state mutation, atomically claim and sync them before apply, retain failed claims, and never delete replacement requests. Keep internal journal reconciliation and legacy qualification unchanged.
Co-authored-by: GPT-6 <gpt-6@openai.com>
Keep operator-owned native circular extrapolation distinct from cold/adaptive qualification and playback readiness. Preserve qualified frontend/assets, use guarded journal transactions and rollback, and apply private startup requests before the owned monitor begins. Retain legacy rendering and qualification gates.
Co-authored-by: GPT-6 <gpt-6@openai.com>
Require effective SYS_PTRACE only for native owned workers and distinguish denied FD inspection from foreign queue ownership. Keep inode proof authoritative and constrain deployment to its private PID namespace.
Co-authored-by: GPT-6 <gpt-6@openai.com>