StackMap
Subscribe

Acontext vs book-to-skill

Skill memory layer for agents: auto-captures learnings from runs into plain Markdown skill files you can read, edit, git and share across frameworks — memory without an opaque store. — versus — Turns any technical book PDF — or folder of sources — into a structured agent skill: mental models in SKILL.md, per-chapter files loaded on demand. Claims 24-51x fewer tokens than context-dumping.

The curated verdict

Two sources for the same artifact: Acontext distills skills from what your agents DID; book-to-skill distills them from what experts WROTE. Both land as plain skill files.

Acontextbook-to-skill
Stars3.7k18k
Forks3322.0k
LanguageJavaScriptPython
LicenseApache-2.0MIT
Last activity26 days ago2 days ago
Topicsmemory, skillsskills
Curated connections63

Acontext — the curator's take

The "memory should be legible" bet: instead of embeddings in a vector store, learnings from agent runs become Markdown skill files you can read, diff, git and mount into any framework — debuggable memory users can inspect and correct, the exact thing opaque memory layers get wrong. It can also adopt and evolve skills you wrote or downloaded. Trade-off: no semantic recall over thousands of entries; it lives or dies on distilling runs into a curated, manageable skill set. Want scale-out retrieval memory instead? That's memmachine or hindsight territory.

book-to-skill — the curator's take

Distillation over summarization is the insight: it extracts frameworks, decision rules and anti-patterns into a skill your agent loads per-question, so the book answers from real content instead of hallucinated recall — and it targets the open Agent Skills standard, not one vendor. When NOT: copyright is on you — a distilled skill of a purchased book is fine on your machine, not in your repo; and distillation quality tracks the source's structure — sprawling books distill poorly.