153 tagged agent-evaluation, measured the same way as everything else here.
Browse within: agent-orchestration 138agentic-control-plane 125agentic-os 125agentic-workflows 125ai-agent-governance 125benchmark 89llm-agents 85ci 83python-cli 83release-gate 83ai-coding 37coding-agents 26codex-cli 19agent-benchmark 18
Skill Claude CodeCodex ✓ vendor
Use when converting an existing benchmark, rubric, verifier, task YAML/JSON, or domain check into SkillEvaluator BYOG/BYOT custom evaluation.
Skill Claude CodeCodex
Inspect accessible names, roles, and tree structure to drive or audit browser UI behavior.
samarailly51-pixel/claimpilot-harness
Skill Codex
Review and structure auto insurance bodily injury claims, including intake triage, evidence completeness, injury causation, treatment chronology, medical necessity, wage-loss support, negotiation risks, and human escalation. Use when analyzing bodily injury claim files, designing claims-agent workflows, drafting…
Skill Claude Code
Project code-review style for this repo. Review the current branch's changes in two passes (correctness + justify-every-line), tracing any bug claim to ground before reporting.
multi-agent-systems-failure-taxonomy/ATLAS
Skill Claude Code
Use when working on an agent task where AdaMAST failure-mode checkpoints, final submission gates, trace capture, taxonomy generation/refinement, or AdaMAST CLI setup should guide Codex. This skill helps Codex apply AdaMAST during software or research tasks, diagnose its own trajectory against the active taxonomy, and…
lexmount/browseruse-agent-bench
Skill Claude CodeCodex
Create and integrate new modular agents for browseruse-bench (BaseAgent in browserusebench/agents/, registry import in init.py, entry in configs/agentregistry.yaml, runtime config in root config.yaml agents section). Use when a user asks to add/register a new agent, scaffold a custom agent, or implement a new agent…
Skill Claude Code
Run coding agent benchmarks and verify results with nasde. Use this skill when the user wants to: Run a benchmark (all tasks, single task, specific variant) Re-run assessment evaluation on existing trial results Check or verify results in Opik (traces, feedback scores, experiments) Troubleshoot a failed benchmark run…
Skill Claude Code
Publish an already validated report to an external release destination.
Skill Claude CodeCodex
Design and scaffold a new trapstreet.run task to evaluate a given agent/skill/tool -- the reverse of trapstreet-solution-scaffold (a solution for an existing task). Generates the mechanical parts (traptask.yaml, judge.py/grader.py on the TRAPTASKMANIFEST contract, buildcases.py's validate-then-render pipeline) and…
nnabuuu/harness-engineering-toolkit
Skill Claude CodeCodex
A codebase scan for five common weaknesses in the setup used to guide and check an AI agent. It produces a diagnostic report and can optionally fix the problems it finds.
lindixu6-hash/ai-content-workflow-skills
Skill Claude CodeCodex
A Chinese-language writing workflow for creating posts and spoken scripts for Xiaohongshu, a Chinese social media platform. It develops a topic through several draft stages and preserves the user’s real experiences and evidence.
Skill Claude CodeCodex
Use when an agent needs to search the shared knowledge graph, contribute a discovery, or update existing entries. Covers search, tagging, discovery types, and status lifecycle.
Skill Claude CodeCodex needs its repo
Author and run Apo tasks — executable capability specs for AI agents. Use when writing an apo adapter or task (.eval.ts), defining checks or input files, publishing a task catalog, or running agent tasks and reading PASS/FAIL verdicts with apo task run / apo runs show. Covers the define, execute, verify, improve loop.
Skill Claude CodeCodex
Keep the Codex ablation example focused on the requested file change.
aaronartistzhang-afk/DailyWork
Skill Claude CodeCodex
A workflow for comparing two versions of a prompt, skill, or piece of text without revealing which one is the new version. Separate agents create the versions, and a person scores them without seeing their labels.
Oshawott324/datalox-gated-runtime
Skill Claude CodeCodex
Operate the local commerce support world using provider evidence and safe shadow writes.
At most 3 mods per repository are shown here, and a mod shipped inside a plugin is left to that plugin's page — the rest are on their repository pages: