Notes on agent selection, the Fold Score, and how coding agents choose and use developer tools.https://nativefold.com/https://nativefold.com/blog/agent-accuracy-vs-reliability/https://nativefold.com/blog/agent-accuracy-vs-reliability/A benchmark score climbs when you buy more retries. Repeatable behavior does not. Why agent accuracy and reliability differ, and must be measured separately.Fri, 17 Jul 2026 00:00:00 GMTagent reliabilitybenchmarksevaluationAI agentshttps://nativefold.com/blog/agent-evaluation-variance/https://nativefold.com/blog/agent-evaluation-variance/Agent evaluation variance is real: across 60,000 trajectories, ten identical benchmark repetitions spanned 2.2 to 6.0 percentage points per configuration.Tue, 14 Jul 2026 00:00:00 GMTagent evaluation variancesingle-run pass@1how many runs to evaluate an agenthttps://nativefold.com/blog/agents-your-analytics-cannot-see/https://nativefold.com/blog/agents-your-analytics-cannot-see/Your analytics can see an SDK was installed, not that an agent selected it. Why agent tool selection is invisible to product analytics, and how it is measured.Wed, 15 Jul 2026 00:00:00 GMTagent analyticsagent tool selectionagent-native market shareproduct analyticshttps://nativefold.com/blog/geo-for-developer-tools/https://nativefold.com/blog/geo-for-developer-tools/GEO for developer tools in 2026 is not brand mentions in ChatGPT but which SDK coding agents actually install. Why agent selection is the metric that matters.Tue, 14 Jul 2026 00:00:00 GMTGEO for developer toolsAEOagent-native market sharesingle-run pass@1https://nativefold.com/blog/agent-evaluation-experimental-science/https://nativefold.com/blog/agent-evaluation-experimental-science/An agent benchmark score does not by itself tell you the agent will behave that way again, unless the choices behind it were fixed before the results appeared.Sat, 25 Jul 2026 00:00:00 GMTagent evaluationpre-registrationexperimental sciencebenchmark reliabilityhttps://nativefold.com/blog/why-ai-agents-write-deprecated-code/https://nativefold.com/blog/why-ai-agents-write-deprecated-code/AI coding agents write deprecated code because they favor patterns common in training data, which lags current libraries. Here is why, and how to reduce it.Wed, 29 Jul 2026 00:00:00 GMTai-coding-agentsdeprecated-codedeveloper-toolscode-generationllm-limitations