Chunk long-context documents and delegate analysis to an rlm-subcall subagent.
npx skills add https://github.com/ivalmart/Skilled-LLMs-CMPM280G --skill rlm-ivalmart
Please help me install this Agent Skill. Skill: rlm Source: https://github.com/ivalmart/Skilled-LLMs-CMPM280G/tree/main/.claude/skills/rlm Command: npx skills add https://github.com/ivalmart/Skilled-LLMs-CMPM280G --skill rlm-ivalmart
Long-context prompts can exceed standard model limits, making thorough analysis difficult. This skill provides a repeatable workflow to chunk content, route analysis to a subagent, and synthesize results in the main chat.
Initialize the REPL with a context file and start the RLM workflow to process your document.
High-intent search queries and answers about installing and using this skill.
To process long-context documents that exceed model limits, use a Recursive Language Model workflow to chunk content, route analysis to a subagent, and synthesize results in the main chat.
Analyzing large codebases or logs with limited context windows involves deterministic chunking and delegating chunk-level analysis to a subagent, which synthesizes findings into a consolidated final answer.
A Recursive Language Model workflow for large documents uses a persistent REPL and local state to manage chunks, buffers, and subagent analysis, ensuring repeatable processing of content that exceeds standard model limits.
Yes, orchestrating long-context tasks with chunking relies on a persistent Python REPL to maintain local state across invocations, manage chunk files, and coordinate subagent analysis.
The best way to synthesize findings from chunked text is to route each chunk to a subagent for analysis and then consolidate the individual results into a final answer within the main chat environment.
When using chunking and subagents for long-context processing, the approach requires managing local state and buffers across a persistent REPL, which adds orchestration overhead compared to native long-context models.
Ju-uscOrchestrate long-context processing with a root LLM, persistent REPL, and subagent for chunk analysis.
AllanYiinExtract traceable evidence from long documents via chunked REPL queries.
jdubrayCoordinates a persistent Python REPL and subagent to analyze long documents in chunks.
adrianweddProcess arbitrarily long contexts via recursive language modeling with BM25 search.
