REPORT — oracle-classify-correct-refreeze-reviewer
STEP: oracle-classify-correct-refreeze
REV: 7518ae8   FINAL: 3381145102fcdf67e763314d4ec935af4d9bd47f
CHANGED: docs/research/oracle-classify-correct-refreeze-review.md
VERIFIED: scripts/check-scope 7518ae8 3381145102fcdf67e763314d4ec935af4d9bd47f docs/research/oracle-classify-correct-refreeze-brief.md — scope: ok (2 path(s) in 7518ae8..3381145102fcdf67e763314d4ec935af4d9bd47f), tripped: none
  .venv/bin/python -m pytest tests/test_phase3_project_oracle.py::test_project_oracle_classifies_the_five_case_corpus_on_a_real_repository -p no:cacheprovider -q — 1 passed in 444.25s (0:07:24)
  .venv/bin/python -m pytest tests/test_phase3_project_oracle.py::test_gate_config_and_test_corpus_neutralization_is_rejected -p no:cacheprovider -q — 1 passed in 113.97s (0:01:53)
  red re-drive, scratch copy /tmp/review-red with `git show 7518ae8:src/orchestrator/project_oracle.py` restored (diffed against the FINAL file; no identity code remained): /home/ruben/Projectos/meta-project/orchestrator/.venv/bin/python -m pytest tests/test_phase3_project_oracle.py::test_project_oracle_classifies_the_five_case_corpus_on_a_real_repository tests/test_phase3_project_oracle.py::test_gate_config_and_test_corpus_neutralization_is_rejected -p no:cacheprovider -q — 2 failed in 104.22s; corpus member at src/orchestrator/experiment_phases.py:265 with ExperimentError "calibration case 'known-good' classified 'FALSE_SUCCESS', expected 'CORRECT_ACCEPTANCE'"; neutralisation member at tests/test_phase3_project_oracle.py:444 with AssertionError "assert 'FAIL' == 'PASS'"
  hand nested-gate reproduction (/tmp/repro_nested.py; NAMESPACE_V1, _HIDDEN_ENV, the exact gate argv, and the gate's limits: wall 300s, cpu 600, memory 8 GiB, file 1 GiB, descriptors 1024, processes 256) over the reverted tree: nested `scripts/check-tests --fast -p no:cacheprovider -k "not test_frozen_corpus_branch_coverage_meets_the_floor and not test_total_deadline_bounds_a_real_dripping_reply" --hypothesis-seed=424242 ...` — 1 failed, 864 passed, 4 skipped, 89 deselected in 28.34s; the sole failure is tests/test_hook_env_scrub.py::test_a_riding_hooks_path_cannot_silently_disable_a_real_hook at its :228 assertion with stderr "Author identity unknown ... fatal: unable to auto-detect email address (got 'root@rubenlaptop.(none)')"; with the four pinned vars (/tmp/repro_nested_fixed.py): 865 passed, 4 skipped, 89 deselected in 28.13s
  hand inner reproduction (/tmp/repro_inner.py; NAMESPACE_V1, _HIDDEN_ENV, HOME=/tmp, no identity) of `scripts/check-tests --fast -p no:cacheprovider tests/test_hook_env_scrub.py` — 1 failed, 6 passed; the named assertion at tests/test_hook_env_scrub.py:228 fails with the same "Author identity unknown" stderr; with the four pinned vars (/tmp/repro_inner_fixed.py): 7 passed
  scripts/check-tests — 1186 passed in 1164.95s (0:19:24); the full run includes tests/test_phase3_project_oracle.py ....... [53%] (7 passed) and tests/test_dispatch_failure_split.py .. [21%] (2 passed); the full run was not passed --durations=5
  scripts/check-tests --fast --durations=5 — 1076 passed, 110 deselected in 87.13s (0:01:27); slowest five: 8.39s tests/test_phase2_general_target.py::test_phase2_non_self_repository_runs_end_to_end; 7.07s tests/test_phase2_general_target.py::test_patch_mode_never_batches_review_n7; 6.53s tests/test_review_phase6_supervisor.py::test_the_group_kill_reaches_the_grandchild_before_the_supervisor_exits; 5.78s tests/test_phase2_general_target.py::test_crlf_base_refuses_the_lf_context_reply_review_f1; 5.78s tests/test_phase2_general_target.py::test_unterminated_file_refuses_the_markerless_eof_reply_review_n2
  scripts/check-static — exit 0 (gated BLE001/S110 clean over src/orchestrator; 104 advisory B/SIM notes printed under --exit-zero)
  scripts/check-docs — exit 0 (Table verification passed; 409 Total / 183 Unique / 409 OK / 0 Errors)
  hand driver capture (/tmp/capture_driver.py; formats _gate_check for gate-check-static and gate-check-tests): both drivers contain GIT_AUTHOR_NAME, GIT_AUTHOR_EMAIL, GIT_COMMITTER_NAME, GIT_COMMITTER_EMAIL and "oracle-gate"; _GATE_GIT_IDENTITY is the secret-free literal {"GIT_AUTHOR_NAME": "oracle-gate", "GIT_AUTHOR_EMAIL": "oracle-gate@localhost", "GIT_COMMITTER_NAME": "oracle-gate", "GIT_COMMITTER_EMAIL": "oracle-gate@localhost"}
  hand identity comparison: `git config --show-origin --get user.name` -> gituser and `git config --show-origin --get user.email` -> a@b.c; src/orchestrator/worktree.py:39-42 GIT_ENVIRONMENT -> orchestrator/orchestrator@localhost; the pin -> oracle-gate/oracle-gate@localhost
  hand F3 command check: `git rev-parse --verify 617e74f2b88425317bd84860b73f547846ee0a90` prints the name and exits 0; `git cat-file -t` and `git log -1` both fail with "could not get object info"/"bad object"; `git rev-parse --verify 7518ae8` -> 7518ae86b59597d8051740709cffc21b15824a92; the misprinted full sha is not an object
NOT VERIFIED: the implementer's pre-commit hook counts (1042 passed in 23.68s plus 34 passed in 12.34s) and the planner's per-check OUTCOMES hand line were not independently re-run; the two named acceptance tests, the exact nested argv at REV/FINAL, the full module in the standing battery, and the fast suite were. The implementer's module --durations figures were not reused; this review's own durations are above.
FINDINGS:
  F3 | SEVERITY: P2 | ORIGIN: BRIEF | STATE: fixed
  The detection-D brief's REV text is a misprinted full sha (617e74f2b88425317bd84860b73f547846ee0a90) that resolves to no object; this brief carries the rule-8 correction: its own REV 7518ae8 resolves (git rev-parse --verify 7518ae8 -> 7518ae86b59597d8051740709cffc21b15824a92) and its EXPECTED WALLS cites the retrospective scope invocation at the resolved close-out 617e74f4ecc0c5fcdc6cc77e2d236103e5016dff.
  FIXED: c38347c/771b4fa (the committed brief that carries the corrected lineage; the embedded copy matches the standalone file) — no regression test. Caveat: the inherited text's "git rev-parse --verify fails" wording is inaccurate — rev-parse --verify echoes any syntactically valid 40-hex name and exits 0; git cat-file -t and git log are the checks that fail. The substantive no-object finding and the corrected REV stand.
  F4 | SEVERITY: P2 | ORIGIN: BRIEF | STATE: declined
  The brief's suspected interaction — that src/orchestrator/project_oracle.py's --hypothesis-seed=424242 is unrecognised inside the grading sandbox, so the gate exits 4 and the known-good tree grades FAIL through no fault of its own — is refuted, not a defect. The exact nested argv containing --hypothesis-seed=424242 ran 1 failed, 864 passed (with the identity defect) and 865 passed (with the pin); the red re-drive's gate exit is 1, not 4, and pytest accepted the flag.
  FIXED: no regression test (declined — the lead names no defect; the fix is F1's).
  R1 | SEVERITY: P2 | ORIGIN: CODE | STATE: carried
  _gate_check unconditionally injects _GATE_GIT_IDENTITY into the shared gate driver, so gate-check-static receives the four variables too. The static gate runs no git and reads none of them, so its current verdict is unchanged, but the pin is broader than the nested gate it is for and the static check's probe identity/target binding drift with what is otherwise a tests-gate-only fix.
  FIXED: no regression test — hand capture /tmp/capture_driver.py shows both formatted drivers carry all four vars; next action: parameterise _gate_check and pass the identity only from the gate-check-tests call site.
  R2 | SEVERITY: P2 | ORIGIN: CODE | STATE: carried
  The pin is not the identity the trusted host supplies: the host's git config is gituser <a@b.c>, worktree.GIT_ENVIRONMENT is orchestrator <orchestrator@localhost>, and the pin is oracle-gate <oracle-gate@localhost>. The current suite only needs a valid identity (the nested hook test passes with the pin), but the comment's claim "the identity the trusted host would supply" is false and an identity-sensitive test would see a host/sandbox divergence.
  FIXED: no regression test — hand comparison of `git config --show-origin --get user.name/user.email` and worktree.GIT_ENVIRONMENT; next action: pin the trusted GIT_ENVIRONMENT identity or correct the comment and make the identity requirement explicit.
  R3 | SEVERITY: P2 | ORIGIN: BRIEF | STATE: declined
  The GOAL's "re-frozen ... by the registry meta-test's trusted re-freeze act" names an act that does not exist: build_project_oracle derives the frozen corpus from the target tree at build time (the project_oracle_context fixture calls it), and tests/test_dispatch_failure_split.py is already in the target tree, so no freeze input required an edit. The substantive GOAL — known-good CORRECT_ACCEPTANCE, the corpus covers tests/test_dispatch_failure_split.py, the additive row rejects a fresh module, both named tests green — is met; the mechanism clause is over-specified, not a delivered defect.
  FIXED: no regression test (declined — the outcome is verified by the two acceptance tests and the full module; the wording is the planner's).
SHAPES:
  - The inner identity delta is not a suite member: /tmp/repro_inner.py runs the single hook test inside NAMESPACE_V1 with _HIDDEN_ENV and no identity; it fails at tests/test_hook_env_scrub.py:228, and /tmp/repro_inner_fixed.py with the four vars passes 7/7.
  - The exact full nested argv is not a suite member: /tmp/repro_nested.py and /tmp/repro_nested_fixed.py run the gate's argv and limits over the reverted tree, printing 1 failed, 864 passed, 4 skipped, 89 deselected and 865 passed, 4 skipped, 89 deselected.
  - The static-driver env is not built by the suite: /tmp/capture_driver.py captures both formatted drivers and shows the identity in both; no suite member asserts the static driver's env.
  - The host/trusted identity comparison is not built: host git config, worktree.GIT_ENVIRONMENT, and the pin are three different name/email pairs; no suite member asserts the pin equals either.
  - The F3 command behavior is not built: rev-parse --verify exits 0 for a syntactically valid 40-hex name whose object is absent; git cat-file -t and git log fail. The suite does not depend on this command; the inherited evidence does.
  - The GOAL/freeze mechanism was read from build_project_oracle and project_oracle_context; the suite builds its two consequences (the benign tree includes tests/test_dispatch_failure_split.py and the additive row rejects a fresh module), so that is not an unbuilt shape.
TRIPPED: none
PROPOSED ROWS: none
METHOD FRICTION: The INHERITED rule conflict is the one the implementer reports: a received item fully resolved in this step is neither "still open" nor `none`. The reading taken: F3/F4 are listed below with their closed dispositions, and only R1/R2 are open. The FOCUS is a single line longer than 3,000 characters; the Read tool truncates at 2,000, so the review had to re-read it through `fold` — briefs that an agent must read should be wrapped. The full suite was run without --durations=5; the full-run line discloses that, and the fast-suite run above carries the five. No other deviation.
INHERITED:
  F4 | SEVERITY: P2 | ORIGIN: BRIEF | STATE: declined — the suspected --hypothesis-seed=424242 interaction; refuted, see FINDINGS F4.
  F3 | SEVERITY: P2 | ORIGIN: BRIEF | STATE: fixed — the detection-D REV misprint; corrected by c38347c/771b4fa, see FINDINGS F3.
  R1 | SEVERITY: P2 | ORIGIN: CODE | STATE: carried — see FINDINGS R1.
  R2 | SEVERITY: P2 | ORIGIN: CODE | STATE: carried — see FINDINGS R2.
  No other received item is open; R3 is declined and has no action.
---
REPORT — <phase>-<subject>-<role>
STEP: <implementer brief subject>
REV: <copied from the brief's REV — the step's base, not FINAL's parent>   FINAL: <implementer: sha of your work commit, never the commit that holds this report;
  reviewer: the sha under review, copied from the review brief>
CHANGED: <files, one line each>
VERIFIED: <command, duration, counts>
         Every VERIFIED and ACCEPTANCE claim names the command that actually ran — never a
         substituted interpreter, wrapper, or narrowed form. If the documented form cannot run in
         your environment, record both commands and the documented form's separate check. Every
         pytest line names .venv/bin/python -m pytest (bare pytest lacks pytest-timeout; strict
         config rejects it). A suite run records its slowest five (--durations=5): budgets are set
         from measured per-test walls. A timeout is a symptom; classify it before rerunning —
         hang (a surviving descendant or tree; EAGAIN from accumulated children: a kill-surface
         fix), under-budgeted slow test (a heavy but legitimate draw past a fixture wall budget),
         or contention (xdist, a full shared tmpfs — ENOSPC — or a cgroup: isolation, never a
         value change). A "Timeout >120s" line
         is the global firing; anything else naming a timeout (a CANDIDATE_TIMEOUT grade, a
         wall-budget INFEASIBLE, EAGAIN, ENOSPC) is not it.
NOT VERIFIED: <why, or none>
FINDINGS: none | one block each:
  ID | SEVERITY: P0|P1|P2 | ORIGIN: BRIEF|CODE|DOCUMENT | STATE: fixed|declined|carried
  <root cause, one line>
  FIXED: <node id, command, and the named failing assertion at REV — or "no regression test">
         An error or a collection failure is not red. The assertion has to be the thing that
         failed; a run that merely crashed proves nothing.
  STATE means: fixed (above); declined = not a defect, with the reason; carried = a real defect
  left unfixed, passed on in INHERITED. A defect you will not fix now is carried, never declined.
  An ORIGIN: BRIEF finding is fixed once a later committed brief corrects the text: FIXED: names
  that brief's commit and reads "no regression test". Carried means a future actor has something to
  do with it; at closeout a finding with no possible next action is closed — a defect in an
  already-committed report reads FIXED: recorded in the review that filed it, and an authoring
  lesson from a finished brief goes into record.md's Observations. IDs are stable keys: the F<n>
  sequence is the default, and a prefix must name its origin.
SHAPES: <shapes probed that the suite does not build, or none with a reason>
        Every shape this report names needs a test or an entry here: a hand measurement is not a
        test. A facet claimed unchanged is re-driven at FINAL when your fix touches its surface.
        SHAPES: none claims nothing probed is left unbuilt — never that unbuilt probes were left
        unnamed.
TRIPPED: <copied from check-scope's tripped: line>
PROPOSED ROWS: <ledger rows, or none>
METHOD FRICTION: none | <where you deviated from the method or its templates, misread them, or
  found them silent. A defect in the brief is a FINDING, not friction. This form is exhaustive:
  every field is filled and nothing outside it is added. A departure from the form is friction
  even when it changes no outcome — silence is not compliance.>
INHERITED:
<<every INHERITED item you received that is still open, copied verbatim; then each
  finding you marked carried, with its ID>>
Rule 2 governs the planner's copy of this block. What goes in it is yours: an implementer or
reviewer who received an obligation, finding, entry or trigger carries it forward verbatim, and
adds anything new that is still open. A block reading `none` is correct only when nothing came in
and nothing is open. INHERITED never carries residual rows; they are re-selected from the store per
brief.
