Review code and architectural patterns to produce structured advisory reports.
npx skills add https://github.com/ahostbr/liteharness --skill ls-thinker
Please help me install this Agent Skill. Skill: ls-thinker Source: https://github.com/ahostbr/liteharness/tree/main/liteharness/catalog/skills/ls-thinker Command: npx skills add https://github.com/ahostbr/liteharness --skill ls-thinker
It solves the need for fast, high-signal analysis of codebases, architectures, and design tradeoffs without allowing the agent to modify files.
Register the ls-thinker agent, ask it to analyze the provided repository or design description, and then request a structured report focused on findings, prioritized recommendations, and risks.
High-intent search queries and answers about installing and using this skill.
Read-only architecture review inspects existing code and system context without editing files to provide design critique and risk assessment. The analysis generates structured outputs covering summaries, findings, recommendations, and concerns.
Risk-aware advisory for codebase inspection identifies architectural patterns and design tradeoffs to deliver prioritized recommendations and concerns. It evaluates existing systems in read-only mode to ensure no files are modified during analysis.
Structured design recommendations are produced by reviewing code and architectural patterns to output a consistent report. This report includes a summary, key findings, prioritized recommendations, and identified risks or concerns.
Design critique and codebase inspection are supported across LiteSuite sessions by registering the agent and establishing canvas linkage. This allows analysis requests to be registered, processed in read-only mode, and sent back to the requester.
Codebase inspection requires strict non-edit permissions to ensure the agent performs analysis by reading and inspecting existing code without writing or modifying files. This constraint guarantees safe, read-only investigation of system context.
Read-only architecture analysis is limited to inspecting and reviewing existing code without editing, writing, or modifying files. It cannot implement changes or fix identified risks, focusing solely on delivering advisory outputs and recommendations.
korchasaAnalyze documentation and source code to answer user questions in Autonomous Mode.
CANTAGESTUDIODetect architectural hotspots by analyzing churn, coupling, and layer violations.
ahostbrBootstrap a leader and manage worker sessions via the LiteSuite bridge HTTP API.
hoonsubinReview code and suggest refactoring based on clean code principles.
