Analysis Quality Review

Write Report (Analysis Quality Review) — Review-Only Multi-Mode Grader

This is a protocol skill. It grades reader-facing deliverables and ghost-deck storylines. It does NOT edit them. Each invocation runs a single mode and returns a structured response. The calling skill — which has the substantive context that produced the document — writes the spec itself (using a template this skill provides) and applies fixes itself (using references/fix-patterns.md), then re-invokes this skill to verify the fixes landed.

Input convention. Reader-facing deliverables are typically HTML (combined reports with prose + charts/diagrams inline) at standard+ strictness. Markdown is acceptable for working drafts at low strictness. The rubrics and reviewers grade content (argument structure, prose readability) regardless of file format — HTML chrome is ignored, prose and section structure are what's graded. Machine artifacts (YAML datasets, JSON, machine-readable manifests) should NOT be reviewed — they are audit substrate, not reader-facing.

The skill is built on four agents and one mechanism:

• spec mode (no agent): returns the framework-matched authoring template path. File lookup only.
• Pre-flight (review modes only): artifact-loader (runs once per review audit, builds load_bearing_index.yaml)
• spec-judge mode: spec-judge agent — grades the filled-in authoring spec against the declared structural_framework
• review pass: 1: argument-structure-reviewer — grades document structure against the declared structural_framework
• review pass: 2: readability-reviewer — grades document readability (framework-independent — applies to any prose document)

Structural frameworks — pick one per document

Analytical documents don't all use the same structural skeleton. The skill grades against ONE primary framework per document, declared by the caller via structural_framework. Frameworks are not composable; pick the one whose dominant frame matches the document's intent. (Nested Minto-style checks fire inside rumelt-kernel and issue-tree where they apply — handled by the rubric, not by composing frameworks.)

If the caller omits structural_framework, the skill infers a default from doc_type per references/framework-selection-guide.md. The inference is a soft default — callers should override it when their document's intent doesn't match the default (e.g., a doc_type: brief that is descriptive-not-recommendation should be tagged structural_framework: descriptive).

The recommended end-to-end caller sequence at standard+ strictness:

1. spec        → caller fills in template using subject context
2. spec-judge  → loop until PASS or max_spec_iterations (3)
3. write       → caller generates document using approved spec
4. review p:1  → loop until PASS
5. review p:2  → loop until PASS

Stages 1-2 are the ghost-deck discipline (storyline before content). Stages 4-5 are post-hoc grading of the realized document. Each loop's fix step happens in the calling skill, not here.

Pass 1 is gating for Pass 2 (readability review only runs after structure review is clean).

Why review-only

Analytical briefs carry cross-reference codes (SECRET-02, M1-DTC, CP-4), evidence tags ([V], [C]), framework citations (Phase 4.5), and cited numeric claims. The natural fixer of an analytical brief is the skill that generated it — it has the supporting artifacts in working memory, owns the cross-references, and can update brief + dataset coherently. Routing fixes through a separate fixer agent forced us to reinvent that context through a manifest scaffold, which was awkward and error-prone. Now: this skill grades rigorously and hands the calling skill a precise diagnosis it can act on. See references/load-bearing-architecture.md.

When to run

• An upstream analytical skill (reimagine-industry, analyze-industry, build-company-model, stress-test, build-brand, marketing-plan, /quarterly, /decide) has produced a long-form output and wants it verified before handoff.
• A user explicitly requests "quality review", "structure check", "readability audit" on a file.
• A document is about to be handed to a senior reader (partner, sponsor, investor).

Do NOT run on:

• Short chat replies (<200 words)
• YAML config, JSON, raw data files (machine artifacts — not reader-facing)
• Working scratchpads / markdown drafts in working/ folders (audit substrate — not reader-facing)
• SEO/GEO articles (use seo-review-loop instead)
• Code or technical reference material

The rule of thumb: if a human is expected to READ it (gate decision, final report, executive brief), it goes through this review — and at standard+ strictness it is typically HTML. If it's an audit trail, working scaffolding, or machine-readable data, it does NOT go through review.

Calling contract

Upstream skills invoke this protocol by reading SKILL.md and following the orchestration. Inputs (which fields are required depends on mode):

supporting_artifacts, load_bearing_elements, and intent_summary are REQUIRED at strictness ≥ standard for doc_type in {brief, memo, decision-record, deck}. At strictness: low they are optional. At strictness: high the protocol BLOCKS if they are absent. (spec mode doesn't need the manifest; spec-judge and review modes do.) See references/calling-contract.md.

max_spec_iterations: 3 cap. After 3 spec-judge iterations without PASS, the orchestrator returns VERDICT: ESCALATE — the calling skill should consult a human, not keep iterating. The same cap principle applies to review iterations via the calling skill's own bail discipline (no hard cap on the review loop itself).

Return contracts (mode-dependent)

mode: spec returns a single line — the template path. There is no audit dir, no verdict, no model call:

TEMPLATE_PATH: .claude/skills/analysis-quality-review/references/authoring-templates/<doc_type>.md

mode: spec-judge and mode: review return the same five-line block:

PASS: spec-judge | 1 | 2
VERDICT: PASS | FAIL | BLOCKED | ESCALATE
VIOLATION_REPORT_PATH: <path>
FIX_PATTERNS_PATH: .claude/skills/analysis-quality-review/references/fix-patterns.md
AUDIT_DIR: <path>

(For spec-judge, the PASS field reads spec-judge rather than 1 or 2 — it identifies which mode was just graded.) The VERDICT is the single verdict for the pass that was just run; the caller separately tracks whether earlier stages have already cleared.

Orchestration

The orchestrator branches on mode. The three branches share the same five-line return contract (except spec which returns a single template path line).

mode: spec — return authoring template

SPEC ORCHESTRATION (no pre-flight, no model call):
  - Validate inputs: doc_type is set; mode is spec
  - If structural_framework is supplied: use it directly
    Else: infer from doc_type per references/framework-selection-guide.md
  - If doc_type is daily-note: refuse — spec mode does not apply
    (return TEMPLATE_PATH: BLOCKED — see Escalation cases)
  - If structural_framework is concept-page and doc_type is brief/memo/deck: warn but proceed
    (caller has chosen non-default; that's allowed)
  - Resolve to references/authoring-templates/<structural_framework>.md
  - Return single line: TEMPLATE_PATH: <path>

No audit dir is created. No artifact-loader runs. The calling skill consumes the framework-specific template, fills it in using its subject context, writes the filled-in spec to a path of its choosing, and either ships the spec to spec-judge or proceeds without judgment (low-strictness only).

mode: spec-judge — grade the filled-in spec

SPEC-JUDGE PRE-FLIGHT (first invocation only):
  - Validate calling-contract inputs: spec_path, doc_type, intent_summary, supporting_artifacts, template_path
  - If audit_dir not supplied: create tasks/analysis-quality-review/<doc-slug>-<YYYYMMDD-HHMM>/
  - Snapshot the spec to <audit_dir>/00-spec-original.md (first invocation only)
  - NO artifact-loader (the spec has no codes-in-prose yet; cross-references are checked lightly inside the judge)

SPEC-JUDGE PASS (single invocation):
  - If iteration > max_spec_iterations (3): return VERDICT: ESCALATE
  - Invoke spec-judge agent with the manifest + spec + template + previous_violations (if re-invocation)
  - Judge returns verdict + structured report with preservation_note + suggested_fix_shape on each violation
  - Orchestrator returns five-line block

CALLER LOOP (outside this skill):
  - Caller reads report + fix-patterns.md (spec-stage fixes section)
  - Caller revises the spec using its subject context
  - Caller re-invokes spec-judge with the SAME audit_dir, the revised spec, and previous_violations
  - Loop until VERDICT = PASS or VERDICT = ESCALATE
  - On PASS: caller proceeds to write the document; the spec-judge audit_dir is RETAINED for traceability but the review modes create their own audit_dir

mode: review — grade the realized document

REVIEW PRE-FLIGHT (first invocation only):
  - Validate calling-contract inputs
  - If audit_dir not supplied: create tasks/analysis-quality-review/<doc-slug>-<YYYYMMDD-HHMM>/
  - Snapshot the input file to <audit_dir>/00-original.md (first invocation only)
  - If audit_dir was supplied (re-invocation): skip snapshot; index already exists
  - Run artifact-loader → produces <audit_dir>/load_bearing_index.yaml (first invocation only)

REVIEW PASS (single pass per invocation):
  - If pass == 1: invoke argument-structure-reviewer
  - If pass == 2: invoke readability-reviewer (only valid if Pass 1 has previously returned PASS for this audit_dir)
  - Reviewer returns structured violation report (with preservation_note + suggested_fix_shape on load-bearing violations)
  - Orchestrator returns five-line block to caller

CALLER LOOP (outside this skill):
  - Caller reads violation report + fix-patterns.md
  - Caller applies targeted Edit calls to the document, updating supporting artifacts as needed
  - Caller re-invokes this skill with the SAME audit_dir and previous_violations
  - Reviewer grades whether fixes landed
  - Loop until VERDICT = PASS (or caller bails)
  - Then caller invokes Pass 2 against the same audit_dir, same loop pattern

Recommended end-to-end caller flow (standard+ strictness)

1. caller → spec(doc_type=brief)                            → TEMPLATE_PATH
2. caller fills template using its subject context          → writes filled-in spec
3. caller → spec-judge(spec_path, manifest, iter=1)         → ITERATE | PASS
   if ITERATE: caller revises spec, re-invokes (iter=2,3)   → PASS or ESCALATE
4. caller writes the document using the approved spec
5. caller → review(document_path, pass=1, manifest)         → ITERATE | PASS
   loop until PASS
6. caller → review(document_path, pass=2)                   → ITERATE | PASS
   loop until PASS

Stages 1-3 are the ghost-deck phase. Stages 4-6 are the post-hoc realization grading. Doing both phases is what saves 60-70% of iteration tokens vs post-hoc review alone — most structural failures (broken D1 governing observation, MECE-overlap, Rumelt goal-not-strategy) get caught at the 2-3K-token spec stage rather than at the 50-100K-token document stage.

Pre-flight (applies to mode: spec-judge and mode: review; mode: spec skips pre-flight)

1. Verify the input file exists and is readable (document_path for review modes; spec_path for spec-judge).
2. Verify the file is text (markdown, plain text) — refuse binaries.
3. Verify required agent files exist:
   • .claude/skills/analysis-quality-review/agents/storyline-judge.md (for mode: spec-judge — agent type is storyline-judge, mode name is spec-judge)
   • .claude/agents/storyline-judge.md (project-level registration pointer — required so Task tool can resolve subagent_type="storyline-judge")
   • .claude/agents/artifact-loader.md (registration pointer)
   • .claude/agents/argument-structure-reviewer.md (registration pointer)
   • .claude/agents/readability-reviewer.md (registration pointer)
   • .claude/skills/analysis-quality-review/agents/artifact-loader.md (for mode: review)
   • .claude/skills/analysis-quality-review/agents/argument-structure-reviewer.md (for mode: review, pass: 1)
   • .claude/skills/analysis-quality-review/agents/readability-reviewer.md (for mode: review, pass: 2)
4. Validate the manifest per the strictness × doc_type matrix in the calling contract. BLOCK if required and missing.
5. Resolve doc_type and strictness against references/applicability-matrix.md to determine which dimensions fire. For mode: spec-judge, BLOCK if doc_type is wiki or daily-note (those types skip ghost-deck review by design).
6. If audit_dir not supplied: create it and snapshot the input file (00-original.md for review modes; 00-spec-original.md for spec-judge).
7. For mode: review only — if audit_dir not supplied: invoke artifact-loader with the manifest. It produces <audit_dir>/load_bearing_index.yaml. Both review passes read this index. If the loader returns BLOCKED, surface to caller and exit. Spec-judge does NOT run artifact-loader.

If doc_type is daily-note or chat-reply at strictness: low, write a one-line note to the audit dir, set verdict to SKIPPED, and exit. Do not over-review trivial documents.

Spec-Judge Pass

iteration_number = supplied iteration (default 1; cap 3)
report_path = "<audit_dir>/spec-iter<N>-report.md"

if iteration_number > 3: return VERDICT: ESCALATE (spec hasn't converged after 3 rounds — consult human)

Task(subagent_type="storyline-judge", prompt=<<<
    spec_path: <spec_path>
    doc_type: <doc_type>
    structural_framework: <minto-pyramid | rumelt-kernel | issue-tree | scqa-only | adr | concept-page | descriptive>
    strictness: <strictness>
    iteration: <iteration_number>
    report_path: <report_path>
    intent_summary: <verbatim from manifest>
    supporting_artifacts: <manifest>
    template_path: <template_path>
    previous_violations: <path or "none">
>>>)

parse output: VERDICT, REPORT, FAILED_DIMENSIONS
return five-line block (PASS field reads "spec-judge")

Same fresh-Task discipline as the review passes: each invocation MUST be a new Task(subagent_type="storyline-judge", ...) call, never SendMessage. The judge evaluates the spec from scratch every iteration; previous_violations is an untrusted hint about what the calling skill attempted.

Pass 1 — Argument Structure Review

iteration_number = (count of existing p1-iter*-report.md in audit_dir) + 1
report_path = "<audit_dir>/p1-iter<N>-report.md"

Task(subagent_type="argument-structure-reviewer", prompt=<<<
    document_path: <document_path>
    doc_type: <doc_type>
    structural_framework: <minto-pyramid | rumelt-kernel | issue-tree | scqa-only | adr | concept-page | descriptive>
    strictness: <strictness>
    iteration: <iteration_number>
    report_path: <report_path>
    load_bearing_index_path: <audit_dir>/load_bearing_index.yaml
    previous_violations: <path or "none">
>>>)

parse output: VERDICT, REPORT, FAILED_DIMENSIONS
return five-line block

Why fresh reviewer Tasks across re-invocations

The structure reviewer is a verifier; carrying memory of prior verdicts biases it toward rubber-stamping. Each invocation MUST be a new Task(subagent_type="argument-structure-reviewer", ...) call — never SendMessage. Across iterations, the reviewer reads previous_violations from disk to know which violations the caller attempted to fix, but evaluates the document from scratch.

Pass 2 — Readability Review

Runs only if the most recent Pass 1 verdict for this audit_dir was PASS.

Same single-pass-per-invocation structure as Pass 1. Output paths use p2-iter<N>- prefix. The reviewer is invoked with load_bearing_index_path: <audit_dir>/load_bearing_index.yaml. The reviewer runs a phased internal check:

• Phase A (cheap): regex/grep checks for banned tokens, internal codes, passive voice patterns, jargon hits, "and" in titles. If Phase A fails with ≥5 violations, the reviewer returns FAIL immediately without running Phase B.
• Phase B (model eval): so-what completion test, specificity-vs-abstraction judgment, code/jargon contextual review.

This staging is enforced inside the reviewer agent's procedure.

Escalation cases

Outputs

On any terminal state, the orchestrator appends to <audit_dir>/_summary.md:

• Pass number, iteration number, report path, verdict
• Counts of violations by dimension and by load-bearing class
• Whether previous_violations was supplied and how many were marked resolved vs re-issued vs deferred

Do NOT write 12-operations/harness/notifications/routine-complete.md from inside this protocol. The caller (upstream skill) owns its own routine-complete notification. This protocol is a subroutine.

Strictness profile

Pass 2 D8 (concept density) and D9 (register fit) grade prose against references/tone-of-voice.md — the positive register target. They fire at standard+ for analytical doc types.

Strictness governs review depth — how many dimensions fire and how deeply the reviewer reads supporting artifacts to produce preservation_notes. It does NOT govern fix depth (the calling skill chooses how much to fix per iteration).

Full dimension definitions: references/rubric-structure.md and references/rubric-readability.md.

Doc-type applicability

Full table: references/applicability-matrix.md.

Gotchas

• Symptom: Pass 2 surfaces dozens of readability nits on a document that has no clear governing observation. Cause: Caller bypassed Pass 1 or Pass 1 was not yet PASS. Fix: Always clear Pass 1 first. The two passes are not parallel — readability fixes on broken structure produce polished noise. The orchestrator will BLOCK a Pass 2 invocation if Pass 1 hasn't cleared for the same audit_dir.
• Symptom: Loop never converges; same constraints fail iteration after iteration. Cause: Calling skill is applying fixes that don't land OR is regressing previous fixes. Fix: At iteration 3, the caller should read the previous_violations field carefully — patterns of re-issuance are usually a sign the fix pattern is being misapplied. Surface to the human author.
• Symptom: Reviewer marks a wiki page or daily note FAIL because section headers are noun labels not action titles. Cause: doc_type was set wrong. Fix: Wiki pages and daily notes use noun-label headers by convention. Set doc_type correctly at intake.
• Symptom: Internal code patterns survive the reviewer's Pass 2 check because they appear in cross-reference tables the reviewer considers narrative. Cause: Reviewer must classify each occurrence by zone (narrative vs table). Fix: The reviewer's code-scan checks zone via load_bearing_index.yaml classification. See references/edge-cases.md.
• Symptom: Skill called from a daily-note routine, runs full Pass 1 + Pass 2, takes forever, produces low-value output. Cause: Strictness defaulted to standard for a low-value document. Fix: Set strictness: low and doc_type: daily-note for routine outputs.

Rules

1. Pass 1 gates Pass 2. The orchestrator BLOCKS a Pass 2 invocation if Pass 1 has not cleared for the same audit_dir.
2. Every reviewer invocation MUST be a fresh Task(...) call. SendMessage continuations are forbidden.
3. This skill never edits the document. The calling skill applies fixes using references/fix-patterns.md.
4. Snapshot the original file before the first Pass 1 review. The snapshot is the recovery anchor if fixes go wrong.
5. The protocol is read-only against .claude/skills/analysis-quality-review/agents/. Do not modify agent definitions during a run.
6. If an agent file is missing, BLOCKED with a clear "missing file: X" message.
7. Run artifact-loader ONCE per audit, in pre-flight. Subsequent re-invocations reuse the index — do not rebuild.
8. Apply doc_type and strictness gating at the reviewer level, not the orchestrator level — the reviewers know their dimensions; the orchestrator passes the context.
9. Do not write routine-complete.md from inside this protocol. The caller owns its notification.
10. The calling skill is accountable for preservation. If it cannot safely apply a fix, it marks DEFERRED in the next invocation's previous_violations payload — it does NOT guess on load-bearing edits.

References

• references/authoring-templates/ — per-framework templates returned by mode: spec (minto-pyramid, rumelt-kernel, issue-tree, scqa-only, adr, concept-page, descriptive)
• references/framework-selection-guide.md — when to use each framework, with default-by-doc_type inference table
• references/spec-judge-rubric.md — mode: spec-judge dimensions, tests, pass/fail criteria (framework-aware)
• references/rubric-structure.md — review pass: 1 dimensions (D1-D8), tests, pass/fail criteria (framework-aware)
• references/rubric-readability.md — review pass: 2 dimensions (D1-D9), tests, pass/fail criteria (framework-independent)
• references/tone-of-voice.md — the positive register target for analytical prose (graded by Pass 2 D8 concept density + D9 register fit); the companion to fix-patterns.md — what to aim AT, not just what to strip. Analytical generators (html-output report archetypes, upstream analytical skills) read this at write time.
• references/applicability-matrix.md — which dimensions fire under which framework (Pass 1) and which doc_type (Pass 2)
• references/load-bearing-architecture.md — why the calling skill is the natural fixer and why the artifact-loader is a shared grounding step
• references/calling-contract.md — full contract spec, including the apply-fixes-yourself pattern and recommended end-to-end caller flow
• references/fix-patterns.md — editorial fix patterns the calling skill applies (canonical fix for each dimension; includes a spec-stage section)
• references/deeper-reading.md — Minto, Rumelt, Meadows, Conn & McLean — the operational mechanics drawn from each
• references/worked-examples.md — pass/fail examples for each dimension + a worked end-to-end caller example
• references/learnings.md — accumulated behavioural feedback (starts empty)
• references/edge-cases.md — factual exceptions and edge cases