docs(mrl): close contamination-v2 exact-head gate

[TheHalfMoon]

Summary

Close the MRL-6 contamination lineage and temporal-canary ledger after MRL-0601 through MRL-0607 are present on canonical main.

This candidate changes only the canonical task ledger:

• closes MRL-0601 — training-example lineage contract
• closes MRL-0602 — exact/near/semantic contamination interfaces
• closes MRL-0603 — teacher/prompt/source transformation bindings
• closes MRL-0604 — benchmark-derived-generation flags
• closes MRL-0605 — temporal-canary manifest
• closes MRL-0606 — R2-compatible sealed canary fixture workflow
• closes MRL-0607 — no canary recycling into training/search
• marks MRL-0699 — contamination-v2 exact-head qualification as the closeout candidate

Canonical base

44fcd1c0b45311195b0da682755c449c8cabd06a

Exact candidate head

a27b080875cf0115df6842d6c276fc9347afe1d8

All check evidence for earlier heads is stale. The branch was created directly from canonical main after PR #311 merged.

Canonical implementation evidence

• MRL-0601 — canonical lineage implementation and subsequent construction-identity hardening
• MRL-0602 / MRL-0603 / MRL-0605 — reconciled contamination, transformation, and temporal-canary identity hardening through PR fix(mrl): reconcile contamination transformation and canary identity #309
• MRL-0604 / MRL-0606 — reconciled benchmark flags and sealed canary fixture workflow through PR feat(mrl): reconcile benchmark flags and sealed canary fixture #310
• MRL-0607 — exact-head qualified fail-closed temporal-canary recycling guard through PR feat(mrl): prohibit temporal canary recycling #311, merged as 44fcd1c0b45311195b0da682755c449c8cabd06a

Exact scope

One ledger-only change: specs/mesc-research-loop-v1/tasks.md, with exactly eight [ ] to [x] transitions.

No source, tests, workflows, dependencies, lockfiles, runtime configuration, model/data assets, training assets, authorization artifacts, or execution outputs are changed.

Required qualification

This exact head must pass fresh evidence for:

• Python 3.11 quality
• Python 3.12 quality
• Ruff lint
• Ruff format check
• strict mypy
• full pytest/coverage
• MRL machine-state drift/manual-edit gate
• medscale check
• CodeQL
• zero unresolved material review threads/findings

Exit semantics

The [x] entries in this PR are closeout candidates only while unmerged. MRL_CONTAMINATION_V2_READY attaches only after this exact head passes the required gates, remains the exact PR head through expected-head protected merge, and is then present on canonical main.

This gate grants no model/data/network/runtime/GPU/inference/training/promotion/deployment/release/clinical authority.

MRL_CONTAMINATION_V2_READY != MRL_REAL_PREFLIGHT_ENTERED

MRL-0800 remains ineligible until MRL-0499 and MRL-0599 are also satisfied together with all other canonical prerequisites and current training/runtime governance.

No force-push, rebase, destructive history rewrite, CI/CodeQL weakening, or governance bypass is used.

---

Summary by cubic

Closes the MRL-6 contamination-lineage and temporal-canary ledger in specs/mesc-research-loop-v1/tasks.md, marking MRL-0601 through MRL-0607 and the MRL-0699 exact-head qualification as complete on canonical main. These were previously open [ ] tasks; they are now [x] closeout candidates, valid only while this exact head remains unmerged.

Scope and exit semantics

• Only the task ledger changes: eight checkbox flips, with no source, test, workflow, or config modifications.
• The [x] entries are closeout candidates until this PR merges with passing gates; MRL_CONTAMINATION_V2_READY attaches only then.
• This gate grants no model, data, network, runtime, inference, training, promotion, deployment, release, or clinical authority.

^{Written for commit a27b080. Summary will update on new commits.}

[qodo-code-review]

ⓘ Qodo reviews are paused because your trial has ended. Ask your workspace admin to add credits to resume reviews. Manage billing

[coderabbitai]

Important

• 🔍 Trigger review

This repository does not receive automatic reviews because it has fewer than 10 stars.

⚙️ Run configuration

Configuration used: defaults

Review profile: CHILL

Plan: Pro Plus

Run ID: 9e8f9eda-21da-475d-a445-4a0537558def

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands.}

[TheHalfMoon]

Final exact-head semantic review for a27b080875cf0115df6842d6c276fc9347afe1d8: PASS.

• canonical base remains 44fcd1c0b45311195b0da682755c449c8cabd06a
• live compare remains behind_by=0
• scope is exactly one ledger file with eight checkbox transitions (+8/-8)
• MRL-0601..0607 implementations are already present on canonical main through their qualified implementation/reconciliation merges
• the MRL-0699 checkbox remains a closeout candidate until expected-head protected canonical merge
• MRL_CONTAMINATION_V2_READY grants no real preflight, model/data/runtime/GPU/training/promotion/deployment/release/clinical authority
• Python 3.11 and 3.12 quality jobs passed Ruff lint/format, strict mypy, full pytest, MRL machine-state drift/manual-edit gate, and medscale check
• CodeQL analyze (python) passed
• unresolved review threads: 0

No material semantic finding remains.