evaluation commands

Helicone/helicone

Command Cursor

🧊 Open source LLM observability platform. One line of code to monitor, evaluate, and experiment. YC W23 🍓.

run-llama/ParseBench

Command Claude Code

Integrate a new document parsing pipeline into ParseBench: $ARGUMENTS.

Hmbown/Hegelion

Command Claude Code

Command "hegelion" from Hmbown/Hegelion, covering /hegelion, routing and autocoding loop.

redhat-community-ai-tools/harness-eval

Command Cursor

Gate the agent setup on corpus-validated rules (gating tier). Fast, no LLM, exits nonzero on any finding. Suitable for CI and pre-commit.

redhat-community-ai-tools/harness-eval

Command Cursor

Full qualitative review of the agent setup. Read every file, evaluate quality, redundancy, and optimization opportunities. Produce KEEP/REVIEW/REMOVE verdicts per component.

redhat-community-ai-tools/harness-eval

Command Cursor

Deep-evaluate a single skill with static analysis and qualitative review, both individually and in context of the full setup.

JuanMarchetto/agent-skills

Command

Quick architecture drift summary — counts drift items by severity without producing a full report.

JuanMarchetto/agent-skills

Command

Generate an ARCHITECTURE.md from scanning the current codebase — shows draft for user approval before saving.

JuanMarchetto/agent-skills

Command

Run a full architecture drift audit — finds architecture docs, scans code, compares against implementation, and produces a detailed drift report.

sattyamjjain/proofloop

Command

Side-by-side delta between two scorecards of the same skill.

sattyamjjain/proofloop

Command

Compare skill scores against ideal benchmarks.

sattyamjjain/proofloop

Command

Evaluate the execution quality of a skill or agent.

unshDee/proofrag

Command

Evaluate a RAG/LLM app — generate a golden set, judge it, and produce a scorecard.

sina-heidariaan/cold-run

Command Claude Code

Vet a rule, principle, or heuristic with the Delta Test v2 and decide if it can become a real Claude skill or is a platitude that should die.

sina-heidariaan/cold-run

Command

Cold-run every skill in a directory against a baseline agent and report which ones change nothing, which make the output worse, and which earn their place.

sina-heidariaan/cold-run

Command

Test whether each skill's description actually gets that skill loaded, by routing written requests against descriptions alone.