Language: English | 日本語
A knowledge cycle for AI agents — agent behavior compounds, human judgment sharpens.
Agent Knowledge Cycle (AKC) is a six-phase growth cycle for persistent AI agents. It turns repeated agent experience into reusable knowledge — skills, rules, and documentation — under human approval, keeping the agent aligned with the operator's evolving intent while spending the scarce resource — human attention and judgment — only where it matters. The cycle changes the human too: operating it sharpens the judgment that steers it. Built for people who operate coding agents or persistent AI harnesses day to day, it runs inside Claude Code or any comparable harness.
Companion paper: Harness Alignment and Harness Drift: Why Intent, Unlike Correctness, Resists Automation — doi:10.5281/zenodo.20578272
Try it first: one cp of the standalone
akc-cycle rules file gives any AI
agent the six-phase behavior — see Install.
The bottleneck has moved. Most agent frameworks optimize the agent side: more tools, memory, context, or automation. AKC starts from the inverse constraint: as agent capability grows, the scarce resource is the human attention and judgment required to steer the loop (ADR-0010). Skills go stale, rules accumulate residency cost, and documentation drifts — every phase of the cycle exists to keep that maintenance from consuming the operator's fixed budget.
Intent alignment, not just correctness. Tests and linters can check whether one output passes a specification; they cannot check whether a changing harness still matches what the operator now means, because intent itself moves as the operator's judgment sharpens through use. AKC calls the configuration-layer version of this activity harness alignment and its failure mode harness drift — the derivation is in ADR-0017 and the companion paper.
The cycle changes the human too. Curate and Promote force the operator to decide what knowledge is worth retaining; Measure then tests whether those decisions changed behavior. Over time the agent becomes more coherent and the human becomes better at judging coherence — agent behavior compounds, human judgment sharpens.
AKC turns experience into durable behavior through six phases: Research filters intake, Extract captures reusable patterns, Curate audits what accumulated, Promote moves selected patterns into behavior-shaping rules, Measure checks whether behavior changed, and Maintain keeps documents and artifacts coherent.
flowchart TD
E[Experience] --> R[Research<br/>signal-first intake]
R --> X[Extract<br/>reusable pattern]
X --> C[Curate<br/>structural + semantic audit]
C --> P[Promote<br/>human-gated rule or skill change]
P --> M[Measure<br/>observable behavior]
M --> T[Maintain<br/>docs and artifact hygiene]
T --> E
| Phase | Current external skill | Purpose |
|---|---|---|
| Research | search-first | Search broadly, intake only signal that can change the next action |
| Extract | learn-eval | Extract reusable session patterns with quality gates |
| Curate | skill-health + skill-stocktake + rules-stocktake | Run structural debt checks before semantic review of skills and always-loaded rules |
| Promote | rules-distill | Turn recurring patterns into durable rules |
| Measure | skill-comply | Test whether agents actually follow skills and rules |
| Maintain | context-sync + repo-asset-stocktake | Keep documentation roles clean, and audit non-code assets for consumers that have vanished |
Three design-pattern skills — when-code-when-llm, code-and-llm-collaboration, signal-first-research — carry the cycle's reusable design judgments. The phase set and phase-to-skill bindings are a mutable snapshot, not AKC's fixed essence (ADR-0019).
The lightest install is the standalone shimo4228/akc-cycle rules file — it gives an AI agent the six-phase behavior without installing any phase skills:
# From a clone of github.com/shimo4228/akc-cycle, copy the rule
# into your agent's rules directory.
cp rules/common/akc-cycle.md ~/.claude/rules/common/akc-cycle.mdAdopt in stages: the rules file alone lets the cycle emerge in ordinary conversation; add the phase skills above when you want guided, step-by-step execution; fork any of it — AKC defines the cycle, not the implementation. Skills are scaffolding meant to dissolve once the cycle is internalized (docs/scaffold-dissolution.md).
| Area | Contents |
|---|---|
| Decision record | ADR catalog in docs/adr/, with permanent gaps at 0001, 0006, and 0007 from the v2.0.0 extraction |
| AI navigation | graph.jsonld for the concept map, llms.txt for routing, llms-full.txt for a self-contained factual reference (includes the nine design principles) |
| Specifications | schemas/episode-log.schema.json, schemas/knowledge.schema.json |
| Reference implementation | examples/minimal_harness/, a dependency-free Python demo of the three memory layers and two-stage distill pipeline |
| Routing map | docs/CODEMAPS/architecture.md, the canonical file-level navigation index |
The bidirectional loop can fail on the human side — ADR-0014 names gate complacency, deskilling, and delegation-feedback divergence — and on the artifact side as harness drift. The two can compound, which is why AKC treats maintenance as a cycle rather than a one-time configuration. AKC makes these risks explicit and keeps the human approval gate as the structural defense; it does not claim to eliminate them.
Harness engineering improves the scaffold so outputs are correct on the first
try; AKC keeps the scaffold aligned with what the operator means as that intent
evolves (ADR-0009,
ADR-0017). AKC's individual
operations overlap prior agent-memory work such as Voyager, Agent Workflow
Memory, ReMe, and MemGPT; its delta is loop ownership — a structural human
approval gate (ADR-0005), bidirectional
judgment growth, and attention-side scarcity. The full citation trail is in
ADR-0013,
ADR-0017, and llms-full.txt.
This architecture was first proposed and implemented by Tatsuya Shimomoto
(@shimo4228, ORCID
0009-0002-6168-4162) in
February 2026, building on
Everything Claude Code (ECC)
by @affaan-m, the baseline harness used in daily
practice. AKC emerged when the author's own added skills and rules grew large
enough that stale skills, contradictory rules, and drifting documentation became
their own maintenance problem. The first five cycle skills were contributed to
ECC between February and March 2026; context-sync was developed independently.
AKC carries two DOIs: the concept DOI 10.5281/zenodo.19200726 (used by the badge) always resolves to the latest version, while each archived release has its own DOI — cite the release DOI below.
If you use or reference AKC, cite the archived release metadata in
CITATION.cff, also available as
codemeta.json:
@software{shimomoto2026akc,
author = {Shimomoto, Tatsuya},
title = {Agent Knowledge Cycle (AKC)},
year = {2026},
version = {2.5.0},
doi = {10.5281/zenodo.21067957},
url = {https://doi.org/10.5281/zenodo.21067957},
note = {A knowledge cycle for AI agents -- agent behavior compounds, human judgment sharpens}
}In text: Shimomoto, T. (2026). Agent Knowledge Cycle (AKC). doi:10.5281/zenodo.21067957.
The research-ecosystem hub is
shimo4228/shimo4228; it carries the
canonical relationship map for the broader set of research lines.
| Repository | Relationship to AKC |
|---|---|
| Contemplative Agent | Upstream engineering substrate for AKC's early ADRs and downstream operational re-implementation of the six-phase cycle |
| Agent Attribution Practice | Sibling genre library; AKC = cycle mechanism, AAP = attribution practice content |
| Authorship Strategy | Downstream research line on how outputs diffuse outside the operator-agent pair |
| Attention, Not Self | Sibling research line federated at the ecosystem level |
| doctrine-corpus | Bilingual judgment-eliciting Q&A corpus that includes AKC as one source line |
| existence-proof | Pre-line working repository complementing Authorship Strategy |
Japanese development notes are on Zenn; English translations are on Dev.to.
MIT