ailibrary/
frameworks
About a hundred prompt frameworks are in circulation. Six have an author you can check.
The rest are conventions with an acronym, and several have two or three incompatible expansions live at the same time. These write-ups sort them by provenance, so you can tell which ones you can safely cite.
4 articles · Reasoning and agents
ACE: Agentic Context Engineering, and the shift from writing prompts to evolving them
Tier D · 3 min · Reasoning and agentsTreat the context as an evolving playbook, not a static string.
Chain of Thought: the baseline everything else is measured against
Tier D · 3 min · Reasoning and agentsLet the model write the steps before the answer. The reference result.
GEPA: reflective prompt evolution, and what makes it expensive
Tier D · 3 min · Reasoning and agentsUse the model to rewrite its own prompt, scored on real examples.
ReAct: the pattern every AI agent is a variation of
Tier D · 3 min · Reasoning and agentsReason, act, observe, repeat. The interleave that made agents work.