Analyze repository behavior and cross-file relationships from read-only evidence.
npx skills add https://github.com/hotaq/Sprite_harmess --skill analyze-hotaq
Please help me install this Agent Skill. Skill: analyze Source: https://github.com/hotaq/Sprite_harmess/tree/main/.codex/skills/analyze Command: npx skills add https://github.com/hotaq/Sprite_harmess --skill analyze-hotaq
This Skill helps you understand what a codebase is doing, why it is behaving a certain way, and which explanation is best supported by the repository itself.
Ask for a read-only analysis of the repository or a specific question, and request a ranked, evidence-based explanation with file references and confidence levels.
High-intent search queries and answers about installing and using this skill.
To analyze a codebase for cross-file behavior, this Skill traces relationships across configs, modules, docs, and tests. It synthesizes repository behavior into ranked explanations, applying causal investigation to explain architecture using read-only evidence.
To find why your repository behaves a certain way, this Skill performs causal investigation by reading multiple files. It separates direct evidence from inference, providing confidence labels and concrete file citations to identify the root cause.
The best way to compare explanations when debugging is ranked synthesis. This Skill orders multiple plausible explanations by evidence strength and confidence, ensuring debugging conclusions stay grounded in concrete file citations.
Yes, you can get concrete file citations when reviewing codebase architecture. The analysis requires explicit evidence-vs-inference separation, grounding all ranked answers and compatibility questions in read-only repository evidence.
Yes, this repository analysis approach works without modifying your codebase. It relies strictly on read-only evidence to investigate cross-file relationships, ensuring architecture review and debugging tasks remain non-destructive.
SailfishcCurate minimal multi-file repository context packs for downstream coding tasks.
RobinNorbergOrchestrate competing tracer hypotheses in Claude team mode to explain observed results.
gkganesh12Audit source code against documentation and test suites for architectural drift.
anderson-0Generate Mermaid architecture diagrams from source code with dependency analysis.
