narrative-topology

Extract semantic relationships from long narratives, architectures, or complex discussions using RDF-style triple notation. Generate adjacency matrices to re...

This skill appears to do what it says: a small Python scanner that finds single-line '<<{Subject, Predicate, Object}.' entries and outputs a subject→object adjacency matrix. It does not access the network or secrets. Before running: 1) Run the script in a dedicated directory (or a copy of your docs) because it recursively reads files under the current working directory and will process any matching extensions. 2) Note that the predicate is ignored by the code — if you need predicate-aware graphs, the script will need modification. 3) The scanner only matches one-line triples that end with '}.', so multi-line or differently formatted triples will be missed. If these limitations are acceptable, the skill is coherent and low-risk; otherwise ask the author for predicate handling, multi-line support, or a narrower file-glob.

Type: OpenClaw Skill Name: narrative-topology Version: 1.0.0 The narrative-topology skill is a utility designed to extract semantic relationships from text files using a custom RDF-style triple syntax. The provided Python script (scanner.py) recursively scans the current directory for specific file extensions (.md, .txt, etc.), parses the triples, and outputs a compressed adjacency matrix to stdout. The code contains no network calls, no data exfiltration logic, and no suspicious execution patterns; its behavior is entirely consistent with its documented purpose.