psyclaw 0.27.23 → 0.28.1

package/vendor/ars/academic-paper-reviewer/references/calibration_mode_protocol.md

@@ -0,0 +1,256 @@
+# Calibration Mode Protocol
+
+**Parent skill**: `academic-paper-reviewer`
+**Mode name**: `calibration`
+**Purpose**: Measure a bounded decision-error profile against a user-adjudicated target set, or obtain a cheaper directional readout. This protocol never turns criterion judgements into an absolute quality score.
+
+## Epistemic status
+
+An ordinary review has `calibration_status: NOT_CALIBRATED`. Its categorical judgements are tied to named criteria and evidence in one manuscript; they do not establish a stable interval scale, paper ranking, acceptance probability, or cross-session reproducibility.
+
+The full tier may produce a candidate empirical target profile only after completing the protocol against the current domain, article type, venue criteria, rubric version, review mode, and exact replay-derived `execution_topology_sha256`. `PROFILE_MEASURED` means only that errors were measured on that bounded target set. It is not a claim of universal calibration and must not be transported to a materially different target or review topology.
+
+**Current application boundary:** candidate profile production and live-review
+profile application are separate operations. Until a closed, hash-bound target
+profile schema plus replay validator is shipped, reviewer seats and the current
+Schema 6 package adapter emit `calibration_status: NOT_CALIBRATED`; a prose
+profile ID or apparent topology match cannot upgrade them. The full-tier report
+may describe its measured candidate profile, but must label it
+`application_status: NOT_WIRED_TO_LIVE_REVIEW`.
+
+The directional tier always reports `calibration_status: NOT_CALIBRATED`. Three single panels provide directional observations, not an error profile.
+
+## Inputs
+
+1. **Tier**: `full` by default. Use `directional` only when explicitly selected.
+2. **Empirical target set**:
+   - Full: 5–20 user-adjudicated papers, preferably 10–15, with at least one acceptable-side and one reject-side label.
+   - Directional: exactly three papers—one `minor_revision`, one `major_revision`, and one extreme anchor (`accept` or `reject`).
+3. **Per-paper fields**:
+   - manuscript path or text;
+   - adjudicated verdict: `accept`, `minor_revision`, `major_revision`, `reject`, or legacy `borderline` in the full tier only;
+   - venue, domain, and article-type context;
+   - optional `per_dimension_gold_judgements`, using the same criterion-bound categories as `quality_rubrics.md`, with the criterion source and adjudication rationale.
+4. **Target-profile identity**: domain, article type, venue criteria/version, rubric version, review mode, target-set identifier, adjudication date, and one exact `execution_topology_sha256` derived from actual completed panel provenance. A prose configuration label or intended route is not a match key.
+5. **Session persistence**: session-only. Do not cache manuscripts or profiles across sessions.
+
+### Gold-label isolation
+
+Gold verdicts, dimension judgements, rationales, and human assessments must not enter field-analyst, reviewer, or synthesizer context. Join them only after a panel verdict is frozen. The substrate plan must also be fixed without consulting gold material.
+
+This isolation applies to transport as well as prompt construction: no gold
+verdict, dimension judgement, rationale, human assessment, or target-set
+outcome may enter a provider payload or determine the substrate plan. The join
+is post-freeze only, and joined material cannot be fed back into another panel
+within the same attempt.
+
+## Process
+
+### Phase 0: Intake
+
+- Full tier: verify 5–20 papers and both sides of the binary decision boundary. Warn that small or clustered target sets yield uncertain, non-transportable estimates. Legacy `borderline` papers are reported separately and excluded from binary metrics.
+- Directional tier: verify exactly the required three-paper composition, one run per paper, and no `borderline` item. State that the result remains `NOT_CALIBRATED`.
+- Never infer the tier from paper count; ask when the user's selection is missing.
+
+### Phase 1: Run panels
+
+**Full tier.** Run 3 or 5 panel replicates per paper (default 5), then derive the panel verdict by majority vote. Within each replicate, the five seats must have distinct recorded invocation-context IDs. The current provenance artifact checks only that within-panel separation; it does not compare IDs against earlier replicates and therefore does not establish cross-replicate freshness or independent error processes. Preserve each seat's criterion-bound judgements and evidence; do not convert them to numbers or average them. Report exact-agreement and disagreement patterns for a dimension only when the matching gold judgement exists.
+
+**Directional tier.** Run one panel per paper with distinct recorded invocation-context IDs among its five seats. Preserve the exact panel verdict and each scoring seat's categorical dimension judgements as emitted. Do not ensemble, manufacture variance, combine labels into a panel score, or rank the papers.
+
+**Cross-model verification and actual provenance.** `ARS_CROSS_MODEL` is
+default-on for calibration mode. Follow
+`shared/cross_model_verification.md` § Calibration transport exception. Before
+any provider call, run its closed calibration data-fence collision preflight
+independently on the raw reviewer-configuration bytes and raw manuscript bytes.
+A collision refuses the entire attempt before transport: do not send either
+payload and do not escape, strip, rewrite, truncate, switch delimiters, or
+silently fall back.
+
+Create an `attempt_id`, lock one `substrate_plan` before any gold material is
+consulted, and use the same transport and substrate for every paper and
+replicate. For every completed panel, build and replay-validate
+`review-panel-provenance/1.0` from actual seat executions. All result-producing
+panels in one calibration attempt MUST have one identical
+`execution_topology_sha256`; every artifact must derive `fresh_context: true`
+under its fixed `within_panel_attempt_only` scope. The builder receives no
+attempt-history ledger: even if invocation-context IDs differ within each
+artifact, it cannot detect reuse across papers or replicates. Every candidate
+profile and readout must disclose that cross-replicate freshness is unverified
+and must not describe the repeated panels as independent. If consent, configuration, or
+non-content transport preflight is unavailable before the attempt begins, lock
+all seats to the primary family and execute the complete schedule on that
+homogeneous plan.
+
+A mid-attempt dispatch or substrate failure, topology mismatch, unknown
+required provenance observation, within-panel context reuse, or invalid provenance
+artifact invalidates the whole attempt. Every completed panel becomes
+diagnostic-only and must not enter any aggregate. The only result-producing
+retry uses a new `attempt_id`, an empty aggregate, and restarts at paper 1 /
+replicate 1 on one homogeneous plan. Never resume the failed attempt, mix
+transports or substrates, or emit a profile or directional readout from an
+incomplete, mixed-substrate, mixed-topology, or provenance-unresolved attempt.
+
+### Phase 2: Full-tier decision-error profile
+
+Map adjudicated and panel verdicts as follows: Accept/Minor Revision are the acceptable side; Major Revision/Reject are the reject side. With positive meaning acceptable:
+
+| Metric | Meaning | Reporting boundary |
+|---|---|---|
+| Balanced accuracy | Mean of sensitivity on the two sides | Point estimate plus bootstrap interval |
+| FNR | Acceptable-side papers judged Major/Reject (over-harsh) | Point estimate plus bootstrap interval |
+| FPR | Reject-side papers judged Accept/Minor (too lenient) | Point estimate plus bootstrap interval |
+| Exact four-label agreement | Exact verdict matches | Count and share, with target-set size |
+
+Do not report AUC: there is no continuous rubric score. Do not infer an ordinal paper ranking from the four editorial labels.
+
+For each dimension with adjudicated gold judgements, report a categorical agreement table and `annotated_n=<n>/<N>, missing=<N-n>`. With no annotated gold for a dimension, report `NOT COMPUTABLE`. Do not assign distances between categories, average the labels, or claim a gold-set-wide dimension result from a subset.
+
+### Phase 2.5: Minor/Major boundary (both tiers)
+
+When both `minor_revision` and `major_revision` gold examples exist, report raw counts:
+
+| Gold \ predicted side | Accept + Minor | Major + Reject |
+|---|---:|---:|
+| Minor Revision | stayed minor-side | harsh crossing |
+| Major Revision | lenient crossing | stayed major-side |
+
+If either side is absent in a full-tier target set, report `NOT ESTIMABLE — target set lacks both sides of the Minor/Major boundary`. Never render an all-zero table as evidence of no confusion.
+
+### Phase 2.6: Directional reporting boundary
+
+The directional readout may report only:
+
+- each paper's exact gold verdict, exact panel verdict, per-seat categorical criterion judgements, and `lenient` / `exact` / `harsh` direction;
+- raw direction counts;
+- the raw Minor/Major boundary cells; and
+- raw low/med/high severity-grounding-risk counts from Phase 3.5.
+
+It must not report balanced accuracy, FNR, FPR, AUC, confidence intervals, stability, per-dimension error rates, numeric baseline comparisons, or any measured-profile claim.
+
+### Phase 3.5: Severity-miscalibration measurement (#215)
+
+Decision-level error does not capture findings whose severity relies on an asserted field norm. Classify the grounding risk of each emitted weakness:
+
+- **`high`**: severity depends on a field norm or core-result boundary, but the reviewer supplied no applicable external grounding;
+- **`med`**: the reviewer named a possible standard but did not establish its applicability;
+- **`low`**: severity does not depend on a field norm, or the applicable norm is externally grounded.
+
+The classifier evaluates whether grounding was supplied, not whether its own model knowledge says the norm is correct. It **MUST NOT** guess norm-correctness from model memory. Measurement labels must be adjudicated against the separately maintained `evals/gold/field_norm_severity` asset; the pointer is an anti-circularity boundary, not permission to expose gold material to a review run. Full tier reports counts and target-set-local shares; directional tier reports raw counts only. This is separate from FNR/FPR and does not create a quality score.
+
+### Phase 4: Outputs
+
+The full-tier report uses this structure:
+
+```text
+# Empirical Target Profile for <Reviewer Instance>
+calibration_status: PROFILE_MEASURED
+application_status: NOT_WIRED_TO_LIVE_REVIEW
+profile_id: <id>
+target_match: <domain; article type; venue criteria/version; rubric version;
+               review mode; execution_topology_sha256>
+calibration_panel_provenance: <ordered normalized_manifest_sha256 values for
+                               every replay-valid measurement panel>
+Gold set: n=<N>
+Runs per paper: <3|5>
+Cross-model: <yes/no; configuration and fallback disclosure>
+
+## Decision-error metrics
+- Balanced accuracy: <estimate and interval>
+- FNR (over-harsh): <estimate and interval>
+- FPR (too lenient): <estimate and interval>
+- Exact four-label agreement: <count>/<N>
+
+## Per-dimension categorical agreement
+| Dimension | Agreement/confusion counts | annotated_n | missing |
+| ... |
+
+## Minor/Major boundary
+<raw cells or NOT ESTIMABLE with reason>
+
+## Severity-grounding risk
+<low/med/high counts and target-set-local shares>
+
+## Scope and transport warning
+This is a measured profile on the identified target set and exact execution
+topology, not universal calibration. A mismatch in any target-match field restores NOT_CALIBRATED. The current run's replay-derived
+execution_topology_sha256 is one such field and must match exactly.
+```
+
+The directional report uses this structure:
+
+```text
+# Directional Calibration Readout for <Reviewer Instance>
+calibration_status: NOT_CALIBRATED
+Tier: directional
+Gold set: n=3
+Runs per paper: 1
+
+| Paper | Gold verdict | Panel verdict | Per-seat categorical judgements | Direction |
+| ... |
+
+<raw direction counts>
+<raw Minor/Major boundary cells>
+<raw severity-grounding-risk counts>
+
+Interpretation: directional observations only; no error profile, calibration,
+stability, score, ranking, or transport claim is available.
+```
+
+### Phase 5: Session disclosure
+
+A subsequent live review currently begins with this header, including after a
+directional or full-tier measurement run:
+
+```text
+> calibration_status: NOT_CALIBRATED
+> application_status: NOT_WIRED_TO_LIVE_REVIEW
+> Live profile application is not implemented in the current release; a
+> candidate profile, profile identifier, or apparent target/topology match
+> cannot upgrade this live review.
+> Criterion judgements are evidence-anchored but have no measured error profile.
+```
+
+The following is a **future application example only** and is not emitted by
+the current Schema 6 adapter:
+
+```text
+> calibration_status: PROFILE_MEASURED
+> profile_id: <id>
+> profile_artifact_sha256: <raw profile digest>
+> execution_topology_sha256: <profile digest, exactly equal to current replay-derived digest>
+> Decision errors were measured on <N> adjudicated papers matching <target>.
+> This bounded profile is not universal calibration; any target mismatch restores
+> NOT_CALIBRATED. See the attached uncertainty intervals and coverage counts.
+```
+
+The current disclosure cannot be hidden or upgraded. Full-tier measurement may
+produce a candidate profile, but a candidate cannot be applied to a live review
+until the closed profile schema and exact-match replay validator exist. A later
+directional result never replaces a full-tier candidate; neither may be reused
+outside its target identity.
+
+## Ensembling and interpretation notes
+
+- Full-tier repeats require within-panel context separation and use majority voting for the final categorical verdict. Cross-replicate context freshness is not mechanically verified, so reports label that limitation and never infer independent repeated error processes. Per-dimension categorical agreement is reported as counts, not averaged labels.
+- Directional tier is an explicit one-run exception and cannot report stability.
+- Same-family evaluation can understate error. Cross-model evaluation provides stronger evidence when consented and available, but does not prove evaluator independence or correctness.
+- External studies can motivate hypotheses about leniency or harshness, but their numeric gaps must not be imported as correction factors, thresholds, or target-profile measurements.
+- Neither tier predicts performance outside its target identity, detects every within-paper framing failure, or replaces human editorial judgement.
+
+## Integration with existing modes
+
+| Mode | Interaction |
+|---|---|
+| `full` | Ordinary full review remains criterion-bound and defaults to `NOT_CALIBRATED`; calibration runs a separate measurement workflow. |
+| `re-review` | Emit `NOT_CALIBRATED`; re-review decision rules remain unchanged. |
+| `quick` | Emit `NOT_CALIBRATED` and state that reduced scope has no measured live error profile. |
+| `methodology-focus` | Emit `NOT_CALIBRATED`; candidate profile application is not wired. |
+| `guided` | No calibration profile is inferred from dialogue. |
+
+## References
+
+- Lu, C. et al. (2026). Towards end-to-end automation of AI research. *Nature* 651, 914–919. Decision-level validation motivates reporting explicit class conventions and error profiles; its numeric results are not ARS thresholds.
+- Ren et al. (2026). Evaluator-independence and verifiable-subset guidance. Target-set comparison is bounded evidence, not universal calibration.
+- `shared/cross_model_verification.md` — calibration transport and consent rules.
+- `quality_rubrics.md` — criterion-bound judgement contract.

package/vendor/ars/academic-paper-reviewer/references/changelog.md

@@ -0,0 +1,10 @@
+# Changelog
+
+| Version | Date | Changes |
+|---------|------|---------|
+| Unreleased | 2026-08-15 | Contract-honesty hardening (#739/#740): categorical, criterion-bound review outputs; current live packages fixed at `NOT_CALIBRATED` until measured-profile application has a closed replay contract; typed six-axis panel provenance with exact machine handoff and no binary independence claim. No reviewer-correctness or outcome-improvement claim. |
+| 1.4 | 2026-03-08 | Historical release that introduced a numerical rubric and fixed decision mapping (now retired). Current behavior uses criterion-bound narrative judgements; live packages remain `NOT_CALIBRATED`, and measured-profile application is not yet wired. Quick Mode Selection Guide retained. |
+| 1.3 | 2026-03-05 | DA vs R3 role boundaries with explicit responsibility tables; CRITICAL finding criteria with concrete examples; Consensus classification (CONSENSUS-4/3/SPLIT/DA-CRITICAL); Confidence Score weighting rules; Asian & Regional Journals reference (TSSCI + Asia-Pacific + OA options) |
+| 1.2 | 2026-03 | Added statistical reporting standards reference; enhanced methodology_reviewer_agent with statistical reporting adequacy sub-step |
+| 1.1 | 2026-02 | Added Devil's Advocate Reviewer (7th agent), added re-review mode, expanded review team from 4 to 5 |
+| 1.0 | 2026-02 | Initial version: 6 agents, 4 modes, 3-phase workflow |

package/vendor/ars/academic-paper-reviewer/references/editorial_decision_standards.md

@@ -0,0 +1,236 @@
+# Editorial Decision Standards — Criteria for Editorial Decision Making
+
+This document defines the explicit criteria for Accept / Minor Revision / Major Revision / Reject decisions, for use by `eic_agent` and `editorial_synthesizer_agent`.
+
+---
+
+## 0. Decision Authority by Mode
+
+| Mode | Decision engine | Working scale | Output |
+|------|-----------------|---------------|--------|
+| `full` (sprint contract) | Mechanical synthesizer over reviewer contract v2; the matrix below never overrides it | `block/warn/pass` + `block_class` | Accept / Minor Revision / Major Revision / Reject |
+| `methodology-focus` (sprint contract) | Same mechanical engine, scoped to methods + presentation; no venue-fit dimension | same | four-value enum |
+| `full` / `methodology-focus` without a contract | Synthesis Protocol + the qualitative criteria and recommendation matrix in this file | reviewer recommendations | four-value enum |
+| `re-review` | #576 three-gate contract: `re_review_mode_protocol.md` § Decision Derivation, recomputed by `scripts/check_re_review_synthesis.py` | item verdicts (FULLY/PARTIALLY/NOT_ADDRESSED/MADE_WORSE/CANNOT_VERIFY) | Accept / Minor Revision / Major Revision / user_review_required (Reject is not a Stage 3' decision) |
+| `quick` | Journal-Fit Reviewer assessment only; advisory, not an editorial decision | — | signal |
+| `guided` | Issue-list dialogue; no editorial decision letter | — | — |
+| `calibration` | Exact panel verdicts and criterion-bound judgements compared with a user-adjudicated target set; measurement-only | categorical judgements + four-value verdict enum | bounded error profile or directional readout |
+
+Under a sprint contract, the mechanical synthesizer governs. This file's recommendation matrix is the no-contract path and cannot soften, harden, or override a fired contract action. There is no numerical total-score threshold for an editorial decision. In the no-contract path, the synthesizer must explain which anchored criteria and unresolved findings justify the recommendation; strengths in one dimension cannot arithmetically cancel a decision-bearing failure in another.
+
+---
+
+## 1. Decision Categories
+
+### Decision Symmetry and Register Independence (#574 B1)
+
+These principles govern every category below:
+
+- **Symmetric evidence standards.** An Accept conclusion and a Reject conclusion carry the same evidence burden: Accept requires positive, anchored verification that each criterion is met, exactly as Reject requires anchored evidence that criteria failed. Neither direction gets a wider margin of caution.
+- **Decisions follow criteria, not distributions.** Review rigor comes from the venue's actual criteria and article-type expectations (Reviewer Configuration Card), never from acceptance-rate base rates or an expected decision distribution. Base rates describe other papers, not this one; a review round that produces more rejections is not thereby a better round.
+- **Register is independent of severity.** Tone rules (respectful, constructive) govern WORDING only. They never lower a finding's severity or soften a decision, and adversarial or rigor-signaling framing never raises a severity or hardens a decision.
+
+### Accept
+
+**Definition**: The paper can be published without further review.
+
+**Criteria**:
+- Every applicable core criterion is positively verified
+- No unresolved decision-bearing weakness remains
+- At least 3/4 reviewers recommend Accept or Minor Revision
+- No unresolved major academic issues
+
+**Conditions**:
+- May include minor copyediting suggestions
+- May require final formatting adjustments
+- Does not need to be sent for review again
+
+**Typical scenarios**:
+- Paper has undergone multiple revision rounds, all issues resolved
+- First-pass acceptance when the paper genuinely meets every criterion — the decision follows the criteria, never a frequency expectation (#574 B1: no base-rate anchoring, qualitative or numeric)
+
+---
+
+### Minor Revision
+
+**Definition**: The paper is fundamentally acceptable and can be published after limited modifications; typically does not need to be sent for review again after revision.
+
+**Criteria**:
+- The paper is fundamentally acceptable and remaining issues are limited
+- No issue requires restructuring core arguments or methods
+- At least 3/4 reviewers recommend Accept or Minor Revision
+- Issues can be resolved within 2-4 weeks
+- Modifications do not involve restructuring core arguments or methods
+
+**Typical revision items**:
+- Supplementing a small number of references
+- Clarifying certain methodology description details
+- Improving clarity of argumentation
+- Correcting citation format
+- Adding discussion of limitations
+- Adjusting conclusion wording (avoiding overclaiming)
+
+**Response requirements**:
+- Authors must respond to reviewer comments item by item
+- After revision, reviewed by the journal's handling editor (usually not sent for external review again)
+- Revision deadline: 2-4 weeks
+
+---
+
+### Major Revision
+
+**Definition**: The paper has potential but has significant issues, requiring substantial revision followed by re-review.
+
+**Criteria**:
+- One or more material weaknesses require substantial revision
+- The weaknesses are repairable rather than fatal
+- At least 2/4 reviewers recommend Major Revision or better
+- Issues are serious but fixable (not fundamental design flaws)
+- Revision requires 6-8 weeks of work
+
+**Typical revision items**:
+- Re-analyzing data (additional analysis or correcting errors)
+- Substantially rewriting literature review (missing key references)
+- Supplementing additional data collection
+- Reorganizing paper structure
+- Correcting significant methodological flaws
+- Strengthening theoretical framework application
+- Adding robustness checks
+
+**Response requirements**:
+- Authors must write a detailed point-by-point response letter
+- After revision, sent for re-review (may go back to original reviewers or new reviewers)
+- Revision deadline: 6-8 weeks
+- Typically a maximum of 2 rounds of Major Revision allowed
+
+---
+
+### Reject
+
+**Definition**: The paper is not suitable for publication in this journal, even with revision.
+
+**Criteria (meeting any one may trigger Reject consideration)**:
+- A core criterion has a fundamental unfixable failure
+- The paper cannot become suitable for this venue through revision
+- At least 3/4 reviewers recommend Reject
+- Fundamental unfixable issues exist
+
+**Reject subtypes**:
+
+| Subtype | Description | Suggestion |
+|---------|-------------|-----------|
+| **Reject — Out of Scope** | Topic not within journal scope | Recommend more suitable journals |
+| **Reject — Fundamental Flaw** | Fatal flaw in research design | Suggest redesigning the research |
+| **Reject — Insufficient Contribution** | Lacks originality or incremental contribution | Suggest how to strengthen contribution |
+| **Reject — Premature** | Paper not yet mature enough | Suggest specific improvement directions |
+| **Reject — Resubmit Encouraged** | Has potential but needs fundamental restructuring | Provide detailed restructuring suggestions |
+
+**Even with Reject, must**:
+- Affirm genuine merits where they exist — do not manufacture praise to soften the decision (#574 A1/B1)
+- Provide specific improvement suggestions
+- Recommend more suitable journals (if it's a scope issue)
+- Maintain professional, respectful tone
+
+---
+
+## 2. Decision Matrix
+
+### Decision Matrix Based on Reviewer Recommendations
+
+| Journal-Fit Reviewer | R1 | R2 | R3 | -> Recommended Decision |
+|----------------------|----|----|-----|----------------------|
+| Accept | Accept | Accept | Accept | **Accept** |
+| Accept | Accept | Accept | Minor | **Accept** (with suggestions) |
+| Accept | Accept | Minor | Minor | **Minor Revision** |
+| Accept | Minor | Minor | Minor | **Minor Revision** |
+| Minor | Minor | Minor | Minor | **Minor Revision** |
+| Minor | Minor | Minor | Major | **Minor-to-Major** (depends on specific issues) |
+| Minor | Minor | Major | Major | **Major Revision** |
+| Minor | Major | Major | Major | **Major Revision** |
+| Major | Major | Major | Major | **Major Revision** |
+| Major | Major | Major | Reject | **Major Revision** (last chance) |
+| Major | Major | Reject | Reject | **Reject** (resubmit encouraged) |
+| Major | Reject | Reject | Reject | **Reject** |
+| Reject | Reject | Reject | Reject | **Reject** |
+
+### Special Situation Handling
+
+**Split Decision (evenly divided)**:
+- Example: Accept + Accept + Reject + Reject
+- The Journal-Fit Reviewer (or synthesizer) needs to deeply analyze the cause of disagreement
+- Resolve on the arbitration principles (evidence first, expertise first). A genuinely unresolvable split records the dissent and routes the author to respond to it — revision is the vehicle for that response, not a policy of rounding splits toward the stricter verdict (#574 B1)
+- May consider inviting a fifth reviewer
+
+**One Outlier (one unusual opinion)**:
+- Example: Minor + Minor + Minor + Reject
+- Carefully examine the Reject rationale
+- If the rationale is valid and others missed it, escalate to Major Revision
+- If the rationale is insufficient, maintain Minor Revision but mention the opinion in the Decision Letter
+
+---
+
+## 3. Reviewer uncertainty and competence scope
+
+The 1–5 confidence field is retained only as a reviewer's self-reported scope and uncertainty disclosure. It is not a calibrated probability, expertise credential, vote, or mechanical weight. Never ignore, promote, or resolve a recommendation because of this number. Apply the named criterion to anchored manuscript evidence; preserve unresolved conflict and request additional expertise when the evidence or competence boundary requires it.
+
+### Cross-Dimension Decision Impact
+
+Finding severity is transported from the reviewer cards (the Schema 6 enum, `shared/handoff_schemas.md` § Weakness Object) — this table assigns NO severities (#574 A3); it describes how criterion judgements bear on the DECISION only.
+
+| Situation | Decision impact |
+|-----------|-----------------|
+| Methodology owner identifies a fatal design flaw | Even if other dimensions are excellent, Reject is available |
+| Domain owner identifies a repairable major literature omission | Major Revision, require supplementation |
+| Perspective owner identifies a cross-disciplinary relevance failure | Minor or Major Revision, based on its decision impact |
+| Journal-Fit Reviewer identifies a writing-and-structure weakness | Does not by itself invalidate the academic core; require language or structural revision |
+
+---
+
+## 4. Revision Round Policy
+
+### Standard Policy
+
+| Round | Expectation | Handling |
+|-------|-------------|---------|
+| R1 (First revision) | Respond to all reviewer comments | Send for re-review or handling-editor review |
+| R2 (Second revision) | Respond to residual issues | Usually the journal's handling editor makes the final decision |
+| R3 (Third revision) | Very rare, usually only handling formatting | The journal's handling editor makes the final decision |
+
+### Upgrade/Downgrade Rules
+
+- Minor Revision with incomplete revisions -> May escalate to Major Revision
+- Major Revision with excellent revisions -> May downgrade to Minor Revision or Accept
+- Major Revision with insufficient revisions -> May Reject (infinite revision cycles are not encouraged)
+- Beyond 2 rounds of Major Revision -> Strongly recommend Accept or Reject, no further extension
+
+---
+
+## 5. Professional Ethics of Editorial Review
+
+### Reviewer Ethics
+
+1. **Confidentiality**: The review process and paper content are confidential
+2. **Conflict of interest**: Recuse if there is a collaborative or competitive relationship with the author
+3. **Timeliness**: Complete the review within the committed timeframe
+4. **Constructiveness**: Even when recommending Reject, provide constructive feedback
+5. **Impartiality**: No bias based on author's gender, race, institution, or nationality
+6. **No plagiarism**: Do not use unpublished ideas seen during review
+7. **Appropriate language**: Avoid personal attacks, sarcasm, or demeaning language
+
+### Editor Ethics
+
+1. **Fair decision**: Based on academic quality, not influenced by external pressure
+2. **Transparent process**: Decision letter must clearly explain the rationale
+3. **Reasonable deadlines**: Give authors sufficient revision time
+4. **Appeal channel**: Authors have the right to respond to or challenge review comments
+5. **Consistent standards**: Papers of similar quality should receive similar decisions
+
+### Ethical Considerations for Special Situations
+
+| Situation | Ethical Handling |
+|-----------|-----------------|
+| Author is your student/colleague | Must recuse or disclose the relationship |
+| Paper's viewpoint is opposite to yours | Evaluate argument quality, not correctness of position |
+| Paper uses your theory but misunderstands it | May point it out but cannot require citation of your own work |
+| Suspected data fabrication | Report to the real journal's Editor-in-Chief (EIC); the journal initiates its investigation procedure |
+| Paper is similar to your ongoing research | Disclose potential conflict of interest |

package/vendor/ars/academic-paper-reviewer/references/guided_mode_protocol.md

@@ -0,0 +1,34 @@
+# Guided Mode (Socratic Guided Review)
+
+The design philosophy of Guided mode is to **help authors understand the paper's problems themselves**, rather than passively receiving revision instructions.
+
+### How It Works
+
+```
+Phase 0: Normal Field Analysis execution
+Phase 1: Normal execution of 5 reviews (but not all displayed immediately)
+Phase 2: Does not produce full Editorial Decision; enters dialogue mode instead
+```
+
+### Dialogue Flow
+
+1. **Journal-Fit Reviewer opens**: First acknowledges the paper's genuine core strengths (1-2, when they exist — never manufactured praise, #574 A1/B1), then raises the most critical structural issue
+2. **Wait for author response**: Author thinks, responds, or asks questions
+3. **Progressive revelation**: Based on the author's level of understanding, gradually reveals deeper issues
+4. **Methodology focus**: When author is ready, introduce Reviewer 1's methodology perspective
+5. **Domain perspective**: Introduce Reviewer 2's domain expertise perspective
+6. **Cross-disciplinary challenge**: Introduce Reviewer 3's unique perspective
+7. **Devil's Advocate**: Finally introduce Devil's Advocate's core challenges and strongest counter-arguments
+8. **Wrap up**: When all key issues have been discussed, provide a structured Revision Roadmap
+
+### Dialogue Rules
+
+- Each response limited to 200-400 words (avoid information overload)
+- Use more questions, fewer commands ("Do you think this sampling strategy can capture phenomenon X?" rather than "the sampling is flawed")
+- When author's response shows understanding, affirm and move forward
+- When author's response veers off topic, gently guide back to the main point
+- Can ask the author to read a certain reference before continuing discussion
+
+### v3.6.2 sprint contract status
+
+v3.6.2 introduces sprint contracts for `reviewer_full` and `reviewer_methodology_focus` only. A template for this mode will follow in a subsequent patch release. Until then, this mode runs without contract enforcement and retains its pre-v3.6.2 behaviour.

package/vendor/ars/academic-paper-reviewer/references/integration_guide.md

@@ -0,0 +1,15 @@
+# Pipeline Usage Example
+
+```
+User: I want to write a paper about AI in higher education quality assurance, from research to submission
+
+Step 1: deep-research -> Research report
+Step 2: academic-paper -> Paper first draft
+Step 3: integrity check -> 100% verification of references/data
+Step 4: academic-paper-reviewer (full) -> 5 review reports + Revision Roadmap
+Step 5: academic-paper (revision) -> Revised manuscript
+Step 6: academic-paper-reviewer (re-review) -> Verification review
+Step 7: (if needed) academic-paper (revision) -> Second revised manuscript
+Step 8: integrity check (final) -> Final 100% verification
+Step 9: academic-paper (format-convert) -> Final paper
+```

package/vendor/ars/academic-paper-reviewer/references/quality_rubrics.md

@@ -0,0 +1,84 @@
+# Criterion-Bound Judgement Rubric for Academic Paper Review
+
+## Purpose
+
+This rubric defines what each reviewer must examine and how to make the basis of a judgement inspectable. It does **not** provide a calibrated quality score, a paper-ranking scale, or an acceptance probability.
+
+## Calibration status
+
+Every review must declare one of these states:
+
+- `NOT_CALIBRATED` — the required default. Use this whenever there is no empirical target profile matched to the current domain, article type, venue criteria, rubric version, reviewer/model configuration, and review mode.
+- `PROFILE_MEASURED` — reserved for a package-level resolution backed by a replay-valid, hash-bound empirical target profile whose exact target fields and actual completed-panel `execution_topology_sha256` match. Producing or naming a calibration profile does not by itself authorize this status.
+
+Individual reviewer seats always emit `NOT_CALIBRATED`, because the actual
+completed-panel topology (including fallbacks) does not exist until all seats
+finish. The current Schema 6 package adapter also remains
+`NOT_CALIBRATED`; `PROFILE_MEASURED` must not appear in a live review package
+until a closed profile schema and replay validator are shipped. This is an
+honest implementation boundary, not evidence that calibration is impossible.
+
+Never describe either state as proof that judgements are consistent or reproducible across papers, sessions, fields, venues, or model versions. A directional calibration readout is not an empirical target profile and leaves the status `NOT_CALIBRATED`.
+
+## Required judgement form
+
+For every applicable dimension, report:
+
+| Field | Required content |
+|---|---|
+| Criterion source | The target venue criterion, reporting standard, article-type expectation, or reviewer configuration item being applied |
+| Judgement | `EXCEEDS` / `MEETS` / `PARTLY_MEETS` / `DOES_NOT_MEET` / `NOT_ASSESSED` |
+| Evidence anchors | Specific manuscript locations or bounded absence anchors |
+| Rationale | How the cited evidence bears on the named criterion |
+| Uncertainty or scope limit | Missing information, domain dependence, or reviewer limitation |
+| Decision bearing? | Whether this judgement affects the recommendation, with a reason |
+
+These labels are criterion-bound categories, not numbers. Do not convert them into points, weights, a total, a percentage, or a latent ranking. Do not map any total or count of labels mechanically to Accept, Minor Revision, Major Revision, or Reject.
+
+Editorial recommendations instead follow the applicable contract or the qualitative, evidence-anchored rules in `editorial_decision_standards.md`. The recommendation must identify the particular unresolved criteria that make it appropriate.
+
+---
+
+## Dimension 1: Originality
+
+Judge the claimed contribution against the paper's stated field, article type, and target venue. Examine whether the work identifies a defensible gap; distinguishes its theory, method, evidence, or application from relevant prior work; and avoids overstating novelty. Replication and boundary-testing studies can make an original contribution without introducing a new theory.
+
+## Dimension 2: Methodological Rigor
+
+Judge whether the design can answer the stated research question and whether execution and reporting support the inferences made. Apply paradigm- and design-appropriate criteria, including sampling, measurement, validity threats, analysis choices, uncertainty, transparency, and reproducibility where relevant. A standard applies only when its scope matches the paper.
+
+## Dimension 3: Evidence Sufficiency
+
+Judge whether each material claim has evidence of the right type, quality, relevance, and coverage for that claim and field. Consider counter-evidence, triangulation, source provenance, primary versus secondary evidence, and important omissions where relevant.
+
+There is **no universal minimum source count and no universal peer-reviewed-source ratio**. Literature needs vary by field, article type, claim breadth, evidence base, and venue. A review may apply a numeric requirement only when an identified target venue, reporting standard, or protocol explicitly imposes it; cite that authority and do not generalize it beyond its scope.
+
+## Dimension 4: Argument Coherence
+
+Judge whether the problem, gap, research question, method, findings, and implications form a traceable argument. Identify unsupported logical transitions, conclusions that exceed the evidence, unresolved counterarguments, and contradictions. Do not confuse a familiar rhetorical structure with a sound argument.
+
+## Dimension 5: Writing Quality
+
+Judge whether the manuscript communicates its reasoning precisely enough to be reviewed and used. Separate presentation problems from substantive research quality, and do not penalize non-native phrasing when meaning remains clear. Identify only issues that materially affect interpretation, verification, or venue requirements; route copyediting-level points as minor issues.
+
+## Dimension 6: Literature Integration
+
+Judge whether the manuscript identifies and critically integrates the literature needed to establish its question, conceptual lineage, alternatives, and contribution. Coverage is assessed relative to the claims and field, not by a fixed number of references. Missing work should be named or bounded by a clearly described literature area whenever possible.
+
+## Dimension 7: Significance and Impact
+
+Judge whether the claimed theoretical, empirical, practical, or policy implications follow from the evidence and matter for the stated audience. Separate demonstrated significance from speculative future impact; do not require cross-field reach when the target criterion values a focused contribution.
+
+---
+
+## Narrative synthesis
+
+The synthesis must preserve disagreements and non-compensatory weaknesses. A strong judgement on one criterion cannot numerically cancel a failure on another. Report:
+
+1. criteria positively verified;
+2. unresolved decision-bearing criteria, with evidence anchors;
+3. repairability and the work needed to satisfy each criterion;
+4. material uncertainty or reviewer-scope limitations; and
+5. the resulting recommendation under the applicable decision standard.
+
+If the evidence does not support a judgement, use `NOT_ASSESSED` or state the uncertainty. Do not manufacture precision by selecting a midpoint or averaging reviewers.