# ace-agent/ace

Evolve your language agent with Agentic Context Engineering (ACE)

Repository: https://github.com/ace-agent/ace
Canonical: https://ross.abutalabs.com/products/ace-agent-ace
Language: Python
License: Apache-2.0
License Family: permissive
Last push: 2026-08-24T18:50:00+00:00

## Health v2 (maintenance only)
Score: 61/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 99, release rhythm 35, longevity 20
- inputs: {"age_days": 290, "days_push": 9, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1275, forks 164 (observed 2026-08-28T04:04:12.983993+00:00)

## What it is
ACE is a Python framework for Agentic Context Engineering that lets language models self-improve by treating contexts as evolving playbooks. It uses a Generator, Reflector, and Curator loop with incremental delta updates to accumulate and refine domain-specific strategies.

## Use cases
- improve llm agent accuracy without fine-tuning
- evolve prompts automatically from execution feedback
- adapt a language model to a domain without labeled data
- reduce context collapse in long-running agents
- build self-improving agents that learn from experience

## When to choose
- you want agents that accumulate and refine strategies over time
- you need adaptation without labeled supervision or gradient training
- you want lower latency and cost than fine-tuning or heavy rollout methods

## When to avoid
- you need a simple one-shot prompt with no learning loop
- your task lacks any execution feedback signal for reflection
- you require guaranteed deterministic behavior from your agent

## Facets
- artifact type: framework
- maturity: active
- function: agent-framework, llm-training, prompt-engineering, machine-learning
- domain: large-language-models, machine-learning
- platform: python, cross-platform
- tags: context-engineering, self-improving-agents, llm-adaptation, evolving-playbooks, generator-reflector-curator, ai-agents, retrieval-augmented-generation

## Member repositories
- ace-agent/ace (main) score 61

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:04:12.983993+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T05:03:00.427317+00:00, confidence not recorded.
  - readme: https://github.com/ace-agent/ace (fetched 2026-08-28T04:04:12.983993+00:00, sha 23681c1e396e)
- Data as of 2026-08-30T08:39:29.467469+00:00.