Metraipolis › archive › dc_3dd83d201bc4

Learning by Doing: Autonomously Compiling and Curating Agent Skills from Verified Traces — arXiv:2611.01386

Source: arXiv (en)
Original: https://arxiv.org/abs/2611.01386
License (as archived): CC0-1.0
Archived: 2026-11-03T10:52:36.206Z
SHA-256 of the archived text: e138926f20517f30c06cd2d7ca897bbf6422dcf4003318d5fb60072c8f0e26df
Stable URL of this page: https://metraipolis.com/documento/dc_3dd83d201bc4

How to cite

Learning by Doing: Autonomously Compiling and Curating Agent Skills from Verified Traces — arXiv:2611.01386 — arXiv (en), https://arxiv.org/abs/2611.01386 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_3dd83d201bc4 (Metraipolis archive dc_3dd83d201bc4, sha256 e138926f20517f30c06cd2d7ca897bbf6422dcf4003318d5fb60072c8f0e26df)

BibTeX

@misc{dc3dd83d201bc4,
  title = {Learning by Doing: Autonomously Compiling and Curating Agent Skills from Verified Traces — arXiv:2611.01386},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2611.01386},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_3dd83d201bc4; sha256 e138926f20517f30c06cd2d7ca897bbf6422dcf4003318d5fb60072c8f0e26df},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2611.01386. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Memory systems for production LLM agents mostly store semantic memory (facts about users and domains) or episodic memory (earlier interactions). Research systems such as Trace2Skill, WikiSkill, and SkillRevise learn procedures from agent runs, package them as Skills, and evaluate them on benchmarks. We describe Zep Agent Skills, a production service that gives deployed agents this procedural memory. The service uses only verified attempts as evidence, cites the supporting events for each field of a Skill, and governs the admission and distribution of each Skill version. The service records each task attempt as a Trace with ordered events, an outcome, and a verification strength. A compiler with separate Reflector and Curator roles turns verified Traces into versioned Skills. Scans and admission gates control which candidates become active, and verified use outcomes change the search rank of each Skill. The Agent Skills format is an open standard that Anthropic originated and donated to the Agentic AI Foundation, so any host of the format can load a compiled Skill. Agents retrieve Skills through lexical search with progressive disclosure, over an API or MCP. We evaluate Zep Agent Skills on $\tau^3$-bench banking, AppWorld, and EnterpriseOps-Gym. Skills compiled from the Traces of a teacher model increase the pass rate of three weaker executor models on the same tasks, at every reasoning effort, by 4.5 to 18.8 points. Most of the gain is on the tasks that the teacher solved. On ALFWorld, Skills compiled from the runs of a small model give a gain similar to or larger than the gain that prior work reports for continuous fine-tuning, with no training.

Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.AI, nuovi lavori del giorno).
Categorie arXiv: Artificial Intelligence (cs.AI) ; Computation and Language (cs.CL).
Pagina del lavoro: https://arxiv.org/abs/2611.01386 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito).
Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina.
Letto il 2026-11-03T10:49:38.302Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).