Generate token-efficient AST summaries of source code using Tree-sitter.
npx skills add https://github.com/gaberger/hex --skill hex-summarize
Please help me install this Agent Skill. Skill: hex-summarize Source: https://github.com/gaberger/hex/tree/main/.claude/skills/hex-summarize Command: npx skills add https://github.com/gaberger/hex --skill hex-summarize
Generate token-efficient AST summaries of source files to keep AI context manageable during coding, review, and navigation of large codebases.
Summarize the target path to generate a token-efficient AST overview for all source files.
High-intent search queries and answers about installing and using this skill.
To reduce source code context size for LLMs, generate token-efficient AST summaries using Tree-sitter. This extracts structural code information at configurable detail levels, keeping context manageable for large codebases.
A Tree-sitter AST summary is a condensed, token-efficient representation of source code structure. It helps code exploration by providing configurable detail levels (L0-L3) that allow quick navigation of large codebases without loading full file contents.
Summarize a large codebase for incremental documentation by targeting a specific file path and generating AST summaries. Configure output formats like text, markdown, or json, and set a token budget to manage documentation scope.
Yes, you can set a token budget to ensure code summaries fit within LLM context limits. The summarization process respects your specified budget while applying filters to generate token-efficient AST overviews of target source paths.
Tree-sitter parsing supports different output formats for code context provisioning, including text, json, and markdown. You can specify the desired format alongside target path and filters to generate AST summaries tailored to your workflow.
You should not use AST summaries for AI-assisted coding when the LLM requires exact line-level syntax or raw character data, as summarization abstracts structural details to achieve token efficiency rather than preserving full source fidelity.
hafley66Explain tree-sitter's parsing model and node classification for CST vs AST debugging.
Till-Crazy-Tears-Us-ApartGenerate semantic summaries of Python code using AST and LLM analysis.
guyowenNavigate TypeScript codebases with semantic tools for type resolution and call tracing.
closedloop-aiCompress artifacts into token-budgeted summaries with type-specific strategies.
