Phase 2

Phase 2: summary adequacy and dynamic metadata reclamation

Scientific protocol 2.2; backlog revision 2.2.1. The governing specification is
the frozen protocol. Implementation and evidence are tracked in
issue 84 and its published backlog inventory.

Scope

  • Preserve the seven legacy strategy meanings and all existing replay formats.
  • Define independent policy scope, usable authority, measured work, immutable
    intent tapes and static projection equality.
  • Test the two matched-prefix families and clone controls in 12,600 runs.
  • Prove or explicitly refute F2.1-F2.6 with checked assumptions; construct the
    bounded fixed-point quotients and shortest distinguishing continuations.
  • Implement irreversible handle retirement and the contracted authority
    skeleton, with the separate 1,536-trace descriptive stress campaign.
  • Compare the independent interpreter and the pinned seL4 correspondence subset.
  • Freeze source/claim comparisons, analysis contracts, prospective manuscript
    methods, reproducibility records and the Phase 2 evidence release.

Boundaries

Use complete information. Material carriers, persistence, restart, partial
observation and depth-dependent retention are forbidden. F2.7 and csaI's
material/recovery evidence are deferred to Phase 5. A static projection is
distinct from full replay state. Model work has no wall-clock interpretation.

Exit

Close only after P2-31 and all of its dependencies have accepted evidence.
Require preserved legacy behavior, valid semantic/replay checks, passing
negative controls, checked formal results or explicit refutations, the declared
finite and external validation, complete frozen campaign outputs, and
regenerable analyses and figures. Retain every unsupported or negative result.

Report the scientific outcome as supported, unsupported or
inconclusive, using the frozen inference and claim rules. A positive
divergence effect is not required for completion. A broken oracle, failing
control, missing required validation or unresolved required review cannot
satisfy the exit. Report method correctness, evidence completeness, manuscript
completeness and journal readiness separately.

Phase 2 does not establish whole-kernel refinement, native performance or the
Phase 5 recovery supplement. Journal readiness additionally requires the actual
contribution and independent scholarly review. A research report may state a
valid negative result or an unresolved publication contribution.

No due date
49% Completed
Phase 2: verify paired-workload parent acceptance
All checks were successful
clang-format / check-format (push) Successful in 56s
formal / definitions (push) Successful in 2m6s
markdownlint / markdown-lint (push) Successful in 50s
clang-format / check-format (pull_request) Successful in 58s
formal / definitions (pull_request) Successful in 2m4s
markdownlint / markdown-lint (pull_request) Successful in 29s
ci / smoke (push) Successful in 39m39s
ci / smoke (pull_request) Successful in 40m33s
ci / sanitizers (push) Successful in 1h49m19s
ci / sanitizers (pull_request) Successful in 1h46m39s
phase-2
provenance
tests
workload
#673 opened 2026-10-30 06:38:45 +01:00 by erikinkinen
Phase 2: serialize primary and diagnostic decision ledgers
All checks were successful
clang-format / check-format (push) Successful in 1m3s
clang-format / check-format (pull_request) Successful in 50s
markdownlint / markdown-lint (push) Successful in 28s
markdownlint / markdown-lint (pull_request) Successful in 33s
formal / definitions (push) Successful in 2m14s
formal / definitions (pull_request) Successful in 2m9s
ci / smoke (push) Successful in 39m11s
ci / smoke (pull_request) Successful in 38m7s
ci / sanitizers (push) Successful in 2h5m18s
ci / sanitizers (pull_request) Successful in 2h0m46s
experiment
metrics
phase-2
tests
#672 opened 2026-10-30 06:15:28 +01:00 by erikinkinen
Protocol: accept the declared diagnostic and source inventories
All checks were successful
clang-format / check-format (push) Successful in 1m8s
formal / definitions (push) Successful in 2m15s
markdownlint / markdown-lint (push) Successful in 51s
clang-format / check-format (pull_request) Successful in 1m10s
markdownlint / markdown-lint (pull_request) Successful in 1m2s
formal / definitions (pull_request) Successful in 2m36s
ci / smoke (push) Successful in 40m45s
ci / smoke (pull_request) Successful in 40m33s
ci / sanitizers (push) Successful in 2h15m58s
ci / sanitizers (pull_request) Successful in 2h15m31s
bug
cli
phase-2
tests
#669 opened 2026-10-27 02:21:34 +01:00 by erikinkinen
Phase 2: plan canonical paired diagnostic runs and seed groups
All checks were successful
markdownlint / markdown-lint (pull_request) Successful in 38s
markdownlint / markdown-lint (push) Successful in 39s
clang-format / check-format (pull_request) Successful in 59s
clang-format / check-format (push) Successful in 1m3s
formal / definitions (push) Successful in 2m23s
formal / definitions (pull_request) Successful in 2m25s
ci / smoke (pull_request) Successful in 40m7s
ci / smoke (push) Successful in 41m41s
ci / sanitizers (pull_request) Successful in 2h16m9s
ci / sanitizers (push) Successful in 2h16m30s
cli
experiment
phase-2
tests
#668 opened 2026-10-27 02:05:48 +01:00 by erikinkinen