| name | reading-synthesis |
| description | Rigorous thirteen-part synthesis of a text or talk: deep summary, insights, structure, critique, framework rebuild, and CEO-level takeaways. Triggers: reading synthesis, synthesize this, deep dive, rigorous analysis, deconstruct, book analysis, article analysis, essay breakdown, intellectual synthesis, multi-dimensional analysis, executive summary of ideas, framework extraction. Uses: Read (and related file tools) for attached sources; WebSearch or WebFetch when comparands are missing or context is thin. Outputs: single structured markdown message with fixed section headers per rules/output-sections.md. |
| metadata | {"version":"1.0.0","tags":"reading, synthesis, analysis, nonfiction, fiction, mental-models, critique, executive-summary"} |
Reading synthesis
Turn a supplied title, excerpt, file, or full work into one cohesive, high-value analysis aimed at semi-technical readers who want depth without academic padding.
Non-negotiables
- Use the full rubric in output-sections: produce all thirteen sections with the exact headers and order defined there, unless the user explicitly limits scope in the same turn.
- Ingest the whole artifact when the user provides a path or attachment: read the complete source with file tools before analyzing. If only a fragment is given, analyze that fragment and state what is missing for full-work claims.
- Ban fluff and repetition: no throat-clearing, no duplicate sentences across sections, no vague praise. Every paragraph must earn its place.
- Keep section 8 near 300 words (280–320). Count mentally; tighten if over.
- Deliver exactly ten numbered actions in section 10; each must be specific and schedulable.
- Comparative section (5) requires 3–5 named works or thinkers in the same territory. If the user names none, run WebSearch (or WebFetch on solid results) to propose defensible comparands. Never imply you read books you did not access; if comparison is based on well-known summaries, label uncertainty.
- Label inference: in mindset and decision lenses, separate text-evidenced claims from reasonable inference from tone, structure, or context.
- Format: professional, direct tone; level-2 headers matching the rubric file; optional brief bullets inside sections where density helps.
Process
- Clarify the artifact — title, medium (essay, book chapter, talk transcript, etc.), and whether the user wants the full rubric or a subset. If they are silent, run the full rubric.
- Load content — from message or via Read/Glob as needed. Note length limits: if the source is huge, analyze the portion provided or offer to continue in a follow-up with explicit scope.
- Classify structure — fiction vs nonfiction (or hybrid); pick the appropriate structural template in section 6 of the rubric.
- Research comparands if needed for section 5 — short, targeted queries; prefer stable, citable names and works.
- Draft all thirteen sections in one pass — maintain consistent terminology; cross-reference themes instead of copying paragraphs.
- Self-check before send — headers match rubric; section 8 word band; section 10 has ten items; no empty critique; Socratic Q&A is ordered logically.
Integrated example (abbreviated shape)
User: “Here’s a blog post on compound interest and career capital [paste].”
You: Read the paste → section 6 uses nonfiction outline → section 5 compares to similar career-design writing (names 3–4 real essays/books via search if needed) → section 8 stays ~300 words with no repeat of section 1 → section 9 lists principles as a system → section 10 lists ten actions (e.g. “Schedule 90 minutes weekly for skill stack review”).
Anti-patterns
- Inventing quotations, studies, or comparands that cannot be traced to the source or to named public works.
- Replacing critique with tone policing or generic “some may disagree.”
- Turning the Socratic skeleton into a FAQ unrelated to the work’s through-line.
- Using the CEO section for a full recap of earlier sections instead of stakes, decisions, and behavior change.
