Analyze text into five strata and generate compressibility metrics.
npx skills add https://github.com/vincitamore/misc --skill isda
Please help me install this Agent Skill. Skill: isda Source: https://github.com/vincitamore/misc/tree/main/isda Command: npx skills add https://github.com/vincitamore/misc --skill isda
Irreducible Semantic Density Analysis (ISDA) provides a principled way to assess how much of a text is genuinely necessary to convey its core contribution, by decomposing content into five strata and quantifying compressibility.
Provide a sample text to analyze with a five-strata breakdown and a full metric report.
High-intent search queries and answers about installing and using this skill.
To measure semantic density, you can analyze the text using a five-strata framework that decomposes content into structural, retrievable, derived, curatorial, and novel layers to estimate how much is genuinely necessary to convey the core contribution.
The five-strata framework analyzes text by breaking it down into Structural Skeleton, Retrievable Knowledge, Derived Inferences, Curatorial Decisions, and Generative Novelty to reveal where meaning resides and quantify compressibility.
You can identify unnecessary length and guide edits by applying the five-strata analysis to produce quantitative metrics like SCR, ND, RI, and CBAT, which generate a detailed stratum ledger for evaluating editorial efficiency.
Yes, you can measure originality by isolating the Generative Novelty stratum within the text, which evaluates the layer of generative novelty against other content layers to quantify original contributions.
Yes, semantic density analysis evaluates different layers of content across various text lengths to assess editorial efficiency, measuring the irreducible semantic density to determine if condensation or expansion is needed.
aztr0nutzsClean and format raw text by stripping HTML and normalizing whitespace.
harkersDetect AI writing patterns and rewrite text for human-like tone.
fabioabrReview .txt business drafts to validate content and remove unsupported inferences.
Mahaboob26Format Streamlit text with Markdown, LaTeX, colored text, badges, and icons.
