Repository Issue Activity (beta)

#184 rlm pi-plugin

#180 LocalREPL._temp_cwd: process-wide os.chdir races across concurrent RLM instances (FileNotFoundError, cwd cross-contamination)

#179 Claude models emit XML tool-use syntax instead of ```repl``` fences — code silently never executes, model hallucinates results

#178 AnthropicClient: IndexError on empty response.content in completion()/acompletion()

#166 find_code_blocks() crashes with TypeError when LM response is None

#51 can it be used for Windows OS

#158 Question about availability of RLM trajectory dataset

#113 `pip install rlms` on Python 3.10 silently installs broken v0.0.1a1 stub instead of failing

#157 compress_corpus produces fabricated 'search summary' output instead of compressed code at >100K input tokens (gpt-oss-120b via Cerebras)

#88 Expose standard API to allow usage from third party tools

#148 mistake sorry

#147 Can you share the usage data in the paper?

#146 `llm_query` produces repetitive output with Qwen3-4B — maybe it is better to wrapp prompt in chat template?

#142 would you like a PR for z.ai support?

#28 Benchmark details

#144 openai.BadRequestError: 400 - Invalid JSON body sent to RLM completion endpoint

#31 Kubernetes Pod Sandboxes

#135 Making RLM subqueries run in parallel for faster execution

#123 Inner LLM calls missing usage tracking and response in logs

#117 Vision Language Model Support for Subqueries

#111 Why is metadata placed in an assistant message immediately after system?

#107 Need suggestions on extracting info from a 100 page pdf file

#100 Add azure_anthropic backend for Anthropic models on Azure AI Foundry

#91 Proposals: versioning for responses in persistent multi-turn sessions and an option for a self-modifying system prompt

#97 3 CVEs in visualizer's Next.js and React versions