{"paper":{"title":"Remember Your Trace: Memory-Guided Long-Horizon Agentic Framework for Consistent and Hierarchical Repository-Level Code Documentation","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"MemDocAgent generates consistent hierarchical documentation for entire repositories by planning dependency order and consulting a shared memory of prior work traces.","cross_cats":["cs.CL"],"primary_cat":"cs.SE","authors_text":"Changkyu Choi, Jaehoon Lee, Jee-Hyong Lee, Suyoung Bae, YunSeok Choi","submitted_at":"2026-05-14T08:35:20Z","abstract_excerpt":"Automated code documentation is essential for modern software development, providing the contextual grounding that both human developers and coding agents rely on to navigate large codebases. Existing repository-level approaches process components independently, causing redundant retrieval and conflicting descriptions across documents while producing outputs that lack hierarchical structure. Therefore, we propose MemDocAgent, a long-horizon agentic framework that generates documentation within a single, integrated context spanning the entire repository. It combines two components: (i) Dependen"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"MemDocAgent achieves the best performance over both open and closed-source baselines and demonstrates practical applicability in real software development workflows.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That the Dependency-Aware Traversal Guiding and interactions with RepoMemory will reliably produce consistent and hierarchical documentation without introducing new conflicts or requiring extensive manual correction.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"MemDocAgent generates consistent hierarchical repository-level code documentation by combining dependency-aware traversal with memory-guided agent interactions that accumulate work traces.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"MemDocAgent generates consistent hierarchical documentation for entire repositories by planning dependency order and consulting a shared memory of prior work traces.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"ce51a3ebba7b39233cfdf22c2d2f8bddae71cf3adac5fd3f12eecca6b534965e"},"source":{"id":"2605.14563","kind":"arxiv","version":1},"verdict":{"id":"bc807989-1179-4204-b8eb-c0d7ae0aac81","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-15T01:34:03.872531Z","strongest_claim":"MemDocAgent achieves the best performance over both open and closed-source baselines and demonstrates practical applicability in real software development workflows.","one_line_summary":"MemDocAgent generates consistent hierarchical repository-level code documentation by combining dependency-aware traversal with memory-guided agent interactions that accumulate work traces.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That the Dependency-Aware Traversal Guiding and interactions with RepoMemory will reliably produce consistent and hierarchical documentation without introducing new conflicts or requiring extensive manual correction.","pith_extraction_headline":"MemDocAgent generates consistent hierarchical documentation for entire repositories by planning dependency order and consulting a shared memory of prior work traces."},"references":{"count":91,"sample":[{"doi":"","year":2025,"title":"DocAgent: A multi-agent system for automated code documentation generation","work_id":"bd4961ba-13cc-4969-b569-4a4311b8de9d","ref_index":1,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2025,"title":"CodeWiki: Evaluating AI's Ability to Generate Holistic Documentation for Large-Scale Codebases","work_id":"63d67cc9-9977-469d-b4d1-7f1fd3b4ce84","ref_index":2,"cited_arxiv_id":"2510.24428","is_internal_anchor":true},{"doi":"","year":2024,"title":"RepoAgent: An LLM-powered open-source framework for repository-level code documentation generation","work_id":"807776a6-5142-4381-b717-13c2638f7ee5","ref_index":3,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2010,"title":"Precise documentation: The key to better software","work_id":"d73aca9b-4efb-4e33-9fdc-20dc8fe821d9","ref_index":4,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"10.1016/j.infsof.2014.08.003","year":2015,"title":"Usage and usefulness of technical software documentation: An industrial case study.Information and Software Technology, 57:664–682, 2015","work_id":"975376a6-29b7-415c-8f28-d828b36b96c1","ref_index":5,"cited_arxiv_id":"","is_internal_anchor":false}],"resolved_work":91,"snapshot_sha256":"062244a979b34839f86912e98e72b08b1c362d09437a9edff66b352a64cf1a08","internal_anchors":7},"formal_canon":{"evidence_count":2,"snapshot_sha256":"5a2cad01d62fbe69213a052cfd2bdb6e4042f6a2d2d3059cd92dc7ee4027ac5e"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"}