Tagged evals

Migrate memory v2 concepts into a v3 wiki, with review and retrieval evals

Build evals that measure AI product quality with human, code and LLM judges

Test and harden a skill against golden cases using blind subagent runs

Record skill feedback and iteratively improve skill evals with the skill-up CLI

Create, test, benchmark and optimize agent skills with evals and description tuning

Audit evals, benchmarks and experiments for leakage and self-confirming results

Find remote code execution flaws from OS commands and eval-like sinks in a codebase

Run LLM evaluations on 60+ academic benchmarks such as MMLU, HumanEval and GSM8K

Design agent evals with task specs, Harbor tasks, verifiers and benchmarks

Evaluate agent behavior with Autocontext scenarios and export reusable knowledge

Build and debug RL environments for Atropos training with reward functions and evals

Validate installed skill output against assertion manifests and catch regressions

Build golden-dataset eval suites and regression gates for AI agent releases

Build LLMOps pipelines with model promotion, eval gates and rollbacks

Measure RAG retrieval quality, groundedness and hallucinations in regression tests

Score research ideas on clarity, novelty, validity, feasibility and significance

Scaffold a first eval suite that turns real agent failures into regression cases