About Agentiquette
Agentiquette is the structured, scored, searchable authority on the skills, instruction files, loops, memory systems, and evaluation patterns that make AI agents reliable. Everyone else lists agent resources; Agentiquette evaluates them, with published rubrics, timestamped data, and reproducible benchmarks.
The bottleneck in agents moved from model intelligence to execution infrastructure, and most of the ecosystem hasn't noticed. GitHub stars measure marketing, not reliability. Most published skills are under-specified: no triggers, no boundaries, no tests. Memory without decay handling is a liability. We say so, with evidence, and we publish the rubric we used.
Editorial policy, falsifiably
- Every factual claim is traceable to a primary source or an evidence note.
- Every metric renders with its fetch date or not at all. Metrics older than 30 days hide rather than mislead.
- No score, tier, rank, or editorial verdict has ever been or will ever be for sale.
- Corrections acknowledged within 48 hours; material corrections annotated on-page and logged in the changelog.
- Drafting may be AI-assisted; every published page is human-reviewed by the named editorial desk, which is accountable for it.
The framework we repeat everywhere
Prompts tell agents what to do. Skills teach them how. Loops force them to finish. Memory lets them improve. Evaluation proves they worked. Governance keeps them safe.
If you maintain something we scored and we got a fact wrong, tell us. Corrections make the index better and we log every one.
Contact
Email blocpod@gmail.com. Agentiquette is a blocpod product.