How the Adler Analysis Reading Method Works in Stage 0: A Complete Guide

The Adler analysis reading method is a four-step critical reading process (Structural → Interpretive → Critical → Applicability) that produces a BOOK_OVERVIEW.md file in Stage 0 of the cangjie-skill pipeline, serving as the global context for all downstream skill extraction.

The Adler analysis reading method forms the foundation of Stage 0 (also referred to as "Stage O") in the kangarooking/cangjie-skill repository. This methodology ensures contributors achieve a complete, critical, and actionable understanding of an entire source work before any extraction or skill-generation begins. By rigorously following the workflow defined in methodology/01-stage0-adler.md, the pipeline generates a structured overview that subsequent extractors consume as ground truth.

The Four Steps of the Adler Analysis Reading Method

The method guides a manual reading process through four distinct analytical layers. Each layer builds upon the previous to transform raw text into structured, critical knowledge.

Structural Analysis (结构)

This step identifies the skeleton of the work. The reader must:

• Determine the type of work (methodology, biography, philosophy, etc.).
• Summarize the core thesis in a single sentence.
• List 3–7 first-level arguments and define their relational pattern (parallel, progressive, contrast, or refutation).
• State the core problem the author intends to solve.

Interpretive Analysis (解释)

This step translates the author’s language into the reader’s own words to ensure comprehension:

• Build a terminology glossary recording the author’s personal definitions (minimum 5 terms).
• Re-articulate 5–15 core propositions in plain language.
• Map the argument chain: how each proposition is supported by evidence or logical steps.

Critical Analysis (批判)

The most essential step, Critical analysis prevents downstream skills from inheriting the author’s blind spots. According to the source code in methodology/01-stage0-adler.md, the reader must:

• Identify historical constraints (era-specific assumptions that may be outdated).
• Highlight the author’s positional blind spots (industry, cultural, or personal background limitations).
• Surface unstated assumptions that lack justification.
• Formulate the strongest possible counter-argument.

The conclusions of this step directly populate the Boundary (B) field of every skill derived later, ensuring each skill carries an explicit disclaimer of the source’s limitations.

Applicability Assessment (应用潜力)

A novel addition by the cangjie-skill project, this step bridges analysis and extraction:

• Flag content that can be skill-ified (frameworks, checklists, principles, decision procedures).
• Mark content unsuitable for skill creation (pure anecdotes, raw data, emotional narrative).
• Estimate the number of future skills and prioritize them based on "empowerment for the average person."

Quality Gates and Pipeline Integration

Stage 0 enforces a strict quality gate before the pipeline can advance to Stage 1 (Parallel Extract). The BOOK_OVERVIEW.md must satisfy five checklist items:

1. A single-sentence thesis.
2. A complete skeleton list (arguments and relationships).
3. A minimum of 5 glossary terms.
4. A minimum of 3 critical observations.
5. User-approved overview status.

Pipeline Flow

Once the quality gate passes, Stage 0 produces books/<slug>/BOOK_OVERVIEW.md. Subsequent extractors (e.g., principle-extractor.md, case-extractor.md) read this file to anchor their extraction logic. The skill generation stages (parallel extract, triple verify, etc.) inherit the Boundary data derived from the Critical step, ensuring every generated skill references the source’s explicit limitations.

Rendering the BOOK_OVERVIEW.md Template

The repository uses a Jinja2 template located at templates/BOOK_OVERVIEW.md.template to standardize the Adler analysis output. Below is a minimal Python implementation that renders this template with data gathered from the four analysis steps.

import pathlib
import jinja2

# Load the template (path relative to the repo root)

template_path = pathlib.Path(
    "templates/BOOK_OVERVIEW.md.template"
)
env = jinja2.Environment(
    loader=jinja2.FileSystemLoader(template_path.parent)
)
template = env.get_template(template_path.name)

# Example data collected from the Adler reading

context = {
    "BOOK_TITLE": "Deep Work",
    "AUTHOR": "Cal Newport",
    "YEAR": "2016",
    "BOOK_OVERVIEW.md.template": "book/deep-work/BOOK_OVERVIEW.md",
    "DATE": "2024‑03‑01",
    # Structural

    "Structural_Type": "Methodology",
    "Structural_Thesis": "Focused, distraction‑free work yields massive productivity gains.",
    "Structural_Arguments": [
        "Work deeply → high‑quality output",
        "Shallow work → low‑value tasks",
        "Cultivate habits for depth"
    ],
    "Structural_Relationship": "Progressive",
    "Core_Problem": "How to reclaim concentration in a hyper‑connected world",
    # Interpretive

    "Key_Terms": [
        {"term": "Deep Work", "def": "Professional activities performed in a state of distraction‑free concentration", "diff": "Not just any focused work"}
    ],
    "Core_Propositions": [
        "Deep work is rare but valuable",
        "It can be cultivated with rituals"
    ],
    "Argument_Chain": "Author cites neuroscience, case studies, and productivity experiments.",
    # Critical

    "Era_Limitations": "Pre‑COVID remote‑work era assumptions",
    "Positional_Blindspots": "Tech‑industry bias toward knowledge work",
    "Unproven_Assumptions": "Assumes all knowledge work benefits equally",
    "Strongest_Counter": "Deep work may be less effective for collaborative, creative tasks",
    # Applicability

    "Skill_Candidates": ["Time‑blocking", "Digital‑minimalism"],
    "NonSkill_Content": ["Personal anecdotes"],
    "Skill_Estimate": 4,
    "Priority_List": ["Time‑blocking", "Digital‑minimalism"]
}

# Render the markdown

output_md = template.render(**context)

# Write to the appropriate location

output_path = pathlib.Path("books/deep-work/BOOK_OVERVIEW.md")
output_path.parent.mkdir(parents=True, exist_ok=True)
output_path.write_text(output_md, encoding="utf-8")
print(f"✅ Created {output_path}")

Running this script produces a fully populated BOOK_OVERVIEW.md that satisfies the Stage 0 quality gate and can be handed off to the extraction stage.

Summary

• The Adler analysis reading method is a mandatory four-step process (Structural, Interpretive, Critical, Applicability) applied to every source work in Stage 0 of the cangjie-skill pipeline.
• The Critical step is non-negotiable; it generates the Boundary (B) metadata that prevents bias propagation into downstream skills.
• Output is standardized via the templates/BOOK_OVERVIEW.md.template file and must pass a five-item quality gate before advancing to Stage 1.
• The resulting BOOK_OVERVIEW.md serves as the single source of truth for all extractors in the extractors/ directory.

Frequently Asked Questions

What is the purpose of Stage 0 in the cangjie-skill pipeline?

Stage 0 ensures that contributors develop a complete, critical, and actionable understanding of an entire source work before any automated extraction occurs. It acts as the foundation for the entire pipeline, producing the BOOK_OVERVIEW.md file that all downstream stages reference.

How does the Critical step prevent bias in generated skills?

The Critical step requires identifying the author’s historical constraints, positional blind spots, unstated assumptions, and the strongest counter-argument. These findings populate the Boundary (B) field of every skill, explicitly marking the limitations and context of the source material so skills do not present the author’s views as universal truths.

What are the quality gate requirements for completing Stage 0?

The quality gate requires five specific deliverables: a single-sentence thesis, a complete skeleton list of arguments, at least 5 glossary terms defining key terminology, at least 3 critical observations from the Critical step, and final user approval of the overview document.

Where is the Adler analysis method documented in the repository?

The complete specification resides in methodology/01-stage0-adler.md, with the output template located at templates/BOOK_OVERVIEW.md.template. The high-level pipeline description in SKILL.md references Stage 0 as the mandatory entry point for all skill generation workflows.

{
  "mcpServers": {
    "instagit": {
      "command": "npx",
      "args": ["-y", "instagit@latest"]
    }
  }
}