| name | avoiding-bad-writing |
| description | Use this skill to identify and eliminate deeply problematic semantic, structural, and lexical patterns that undermine high-quality writing. It provides a principle-based diagnostic guide for detecting statistical regression (the "Blurry Sketch" effect), intellectual dishonesty (bias by omission), adjectival inflation (puffery), and formulaic rhetorical contrast. |
Field Guide: Principles of High-Quality Prose
Overview
This skill provides a diagnostic reference for identifying deeply problematic semantic, structural, and lexical patterns in writing. It is based on exhaustive observations of AI-generated content and professional writing standards. It focuses on the behavioral and intent-driven markers of bad writing—such as faking authority, puffing up mundanity, and smoothing over nuance—rather than surface-level syntax or formatting.
Core Behavioral Directive: Explicit Omissions
Always aim for completeness. If for some reason you cannot cover a point that should be explained, make that omission explicit for your readers and collaborators. This prevents the "illusion of completeness" and "bias by omission" common in AI-generated and unencyclopedic prose.
Principles & Diagnostic Guidelines
Use this skill as a primary audit tool for identifying and removing specific "signs" of low-quality writing.
1. Authenticity & Specificity (vs. Statistical Regression)
Identify where unique, nuanced facts are replaced by generic, important-sounding descriptions.
- Audit: Has a specific identity (e.g., a unique historical role) been replaced with a generic accolade (e.g., "revolutionary titan")?
- Reference: See signs.md (Section 1).
2. Grounded Substantiation (vs. Adjectival Inflation)
Identify instances where the text attempts to "force" importance through hyperbolic claims or promotional "peacock" terms.
- Audit: Are arbitrary facts assigned "ongoing significance" or "indelible marks"?
- Audit: Are promotional terms like "legendary," "iconic," or "world-class" used without substantive evidence?
- Reference: See signs.md (Section 2).
3. Intellectual Accountability (vs. Vague Attribution)
Detect phrases that create an "illusion of authority" or consensus while denying the reader the ability to assess the source.
- Audit: Does the text use "experts declare" or "it is believed" to avoid naming specific sources?
- Audit: Are anonymous quantifiers like "many" or "most" used to imply unproven consensus?
- Reference: See signs.md (Section 3).
4. Transparency & Intellectual Honesty (vs. Bias by Omission)
Identify patterns that depart from a disinterested tone or create an artificial "Hierarchy of Fact."
- Audit: Is there an "illusion of completeness"? Are gaps in research hidden rather than made explicit?
- Audit: Is there "False Parity" (giving equal weight to minority views) or segregated "criticism" sections that imply the main text is undisputed?
- Reference: See signs.md (Section 4).
5. Rhetorical Directness (vs. Formulaic Contrast)
Identify structural "tricks" used to appear balanced or deep without substance.
- Audit: Is the text using "Not just X, but also Y" to "retroactively challenge" perceived audience thinking?
- Audit: Are present participle ("-ing") phrases used at sentence ends to inject vague synthesis (e.g., "...highlighting its significance")?
- Reference: See signs.md (Section 5).
6. Narrative Synthesis & Precision
- Trivia Mindset: Is there a collection of disconnected facts or "Mere Appearance" anecdotes without synthesis?
- Circumlocution: Are wordy fillers like "due to the fact that" or "at the present time" present?
- Lexical Selection: Is the text cycling synonyms (protagonist, key player) or overusing high-density AI tokens (Additionally, Delve, Vibrant)?
- Reference: See signs.md (Sections 6-8).
Usage Policy
This skill is strictly descriptive for diagnostics, with one prescriptive behavioral rule: make omissions explicit to maintain completeness and integrity. It focuses on semantic and structural behavior rather than surface-level syntax or formatting.
