Make task qualification count-rate two-sided

[MongLong0214]

Closes #109.

Implements the inclusive 4/6–5/6 comparator count-rate gate, retains task-level exclusions with their measured rates, and removes excluded tasks from the analysis set.

[github-actions]

CommitLore — record lint

Trailers: clean — 1 commit in origin/dev..4d3bab848b43cdab9cc648b02b33220291804781
Active constraints: 11 limits · 21 ruled-out · 9 warnings — from 15 records over 4 changed paths

Active constraints for the paths this PR touches

Limits (11)

• r-qual109 4d3bab8 — the existing M4 qualification rows do not carry reproposal_matches, so they cannot be requalified as count outcomes
• r-rejwork141 5f7af65 — the counts read the diff, so an agent that pursued a rejected approach entirely in conversation and committed nothing registers zero
• r-m5outcome 37f3400 — variance is qualified against a pilot, so an outcome with spread in the pilot and none in the study is still possible; this catches the registration error, not every future one
• r-m4expose df84615 — exposure is observed through the shim's own append, so a harness that never installs the shim is caught by the absent-field refusal rather than by this test
• r-protoreg31 1bd02c1 — the calibration arm cannot be used to select confirmatory tasks, only to size them
• r-exposure52 ba69411 — legacy JSONL artifacts predate model and guard-exposure fields | their absence remains unknown and is never inferred or backfilled
• r-4c73a2 1d82a5b — code inherits the limit of any lexical surface — an alternative implemented under another name is invisible, one named in a string literal is counted; both were true of artifacts
• r-9c2f74 d653153 — the ablation arms cannot discriminate on these fixtures -- no-grade and no-lifecycle are byte-identical to the treatment in 9 of 10 tasks, because the seeds carry one reconstructed record and one task with a lifecycle trailer between them
• r-9c2f74 d653153 — the harness assembles its own projection rather than calling the shipped injector, so what is measured is the harness's rendering of the records, not src/core/inject.ts (issue B-08 · Replace the benchmark harness injector with the actual src/core/inject.ts #36)
• r-5c8b31 60ddc39 — the agent CLI exposes no in-flight turn limit, so a per-task turn budget can only ever be observed with this driver
• r-6c2b95 5da793c — the CLI holds no API key, so any real driver runs through the user's own agent session and cannot be exercised in this environment

Ruled out (21)

• r-qual109 4d3bab8 — treating absent comparator qualification as exclusion | it would hide a missing measurement rather than record a measured rate
• r-rejwork141 5f7af65 — a single weighted rejected-work score | the weights would carry the argument, and a reader whose situation weights dependency additions above hunks cannot recover the components from a composite
• r-m5outcome 37f3400 — reusing matched.length as the count | it equals the re-proposal count only while violations stays zero, so it would measure correctly for the wrong reason and stop the day the other field was fixed
• r-m4expose df84615 — asserting on guard_exposure being present | the M4 failure is exactly a field that looks right while the code behind it does nothing
• r-protoreg31 1bd02c1 — keeping GEE with a small-sample correction | permutation is exact rather than corrected, and the correction literature stops short of an ICC this high
• r-protoreg31 1bd02c1 — a per-task qualification threshold read off outcomes | any such threshold reintroduces the selection that saturated M4
• r-protoreg31 1bd02c1 — leaving the duplicate id and exempting it in the test | the check is right and the history was wrong
• r-exposure52 ba69411 — compute a treatment effect with unknown guard exposure | an old row that never recorded whether the guard reached the run cannot distinguish no treatment from an ignored treatment
• r-8e15c2 70f6bcf — hashing only the files the entry point transitively imports | it is the same judgement call that produced this bug, made with more machinery
• r-4a17d9 68d4c92 — guard versus the control as the primary test | it would confound the route with whether records exist at all, and the interesting question is which route works better once they do
• r-4a17d9 68d4c92 — a blocking guard arm | GUARD-CANNOT-BLOCK measured true and false positives in one score band; the only precision-safe threshold catches one in five
• r-4c73a2 1d82a5b — re-scoring the recorded transcripts and reporting the result | the labels would be chosen with the outcome known, which §4 forbids; the recorded runs were used only to check the surface behaves as intended
• r-4c73a2 1d82a5b — moving violation_if to code as well | it is instrumented-only, and changing an unregistered metric alongside a registered one makes it unclear which change any difference belongs to
• r-6d92c4 b058de2 — documenting grep -a as the convention | it leaves every future reader one forgotten flag away from the same silent empty result, and the cause takes six characters to remove
• r-9c2f74 d653153 — resume the pilot into the same file | a new process would load the edited code and create the mixing that had not happened
• r-9c2f74 d653153 — run the ablation arms as they stand | three nulls from comparing identical inputs read as "these guarantees do not matter"
• r-9c2f74 d653153 — keep the tasks that showed an effect and rewrite only the rest | the property is the criterion, not the direction of the result
• r-5c8b31 60ddc39 — keep turns and explain it in prose | the JSONL outlives the prose, and whoever reads the rows later will not have it
• r-5c8b31 60ddc39 — drop the turn budget since it cannot be enforced | the overrun is still the signal that a run went off the rails
• r-6c2b95 5da793c — skip the dry-run driver | then nothing exercises the harness until a key exists, and the first real run debugs the harness instead of measuring anything
• r-6c2b95 5da793c — emit dry-run rows without a marker | indistinguishable from measurements the moment they leave the terminal

Warnings (9)

• r-qual109 4d3bab8 (claim) — callers must pass the primary comparator's six calibration rows for every candidate task
• r-exposure52 ba69411 (claim) — guard exposure is instrumented in the benchmark hook adapter, not by changing guard scoring | M4 rows remain unexposed and metrics refuses their effect estimate
• r-8e15c2 70f6bcf (claim) — this is the second time in two commits that a fix for the M3 failure did not actually cover the M3 failure. The first was caught by testing the fix against the incident rather than against its own tests
• r-4a17d9 68d4c92 (claim) — both hooked arms pay a subprocess per edit. M2 put all three of its timeouts in its hooked arm, and a timeout truncates a run, which flatters whichever arm suffers it — stated before the numbers exist
• r-4c73a2 1d82a5b (claim) — SURFACES in task-loader.ts and MatchSurface in types.ts are two lists of the same thing — adding a surface means editing both
• r-6d92c4 b058de2 (claim) — keep the NUL separator when touching these keys — a printable delimiter can appear inside a trailer value and would collide
• r-9c2f74 d653153 (claim) — after the measurement, check that git status is clean and the recorded sha is still HEAD -- an edit mid-run breaks reproducibility silently, and that check is the only thing that catches it
• r-5c8b31 60ddc39 (claim) — a driver that gains a real turn limit should stop emitting over-turns and start emitting an enforced label -- do not reuse over-turns for something the harness actually stopped
• r-6c2b95 5da793c (claim) — never cite a row with simulated:true -- README numbers come from bench/results logs and those rows are not results

git log --follow accepts exactly one pathspec, so renames are not followed for 4 paths; query one path at a time to follow its rename chain

withheld the content of 1 record(s) graded blocked: a Ruled-out trailer matching an injection pattern is reported, never quoted (SPEC §7)

_{Trailer violations fail this check. Active constraints are informational — they are what the repository already decided, not a verdict on this PR.}