Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →hashgraph-online/harness-audit
Harness 能力面审计（7 维度评分 + 改进建议）。对角色覆盖、技能完整性、Hook 有效性、规则执行、命令覆盖、文档质量、集成深度进行全面评估，输出 Overall Score、Dimension Scores 和 Top Actions。适用于平台健康检查、季度能力审查、新增能力后的回归验证。关键词：harness-audit、平台审计、技能完整性、Agent 覆盖、命令覆盖、文档质量。
Passed 15 of 19 test cases with the skill, and 4 of 19 without it.
and 15 more in the full report
16 fixed4 passed either way4 still failing1 broke
Tokens, turns and time describe resource use, not a dollar price.
One of 15 test cases the skill fixed. Same model, same prompt, with and without the skill.
During rule enforcement evaluation, how many isolated or unreferenced rules cause the score to land in the 70–89 tier, and how many cause it to drop below 70?
| Checked | Without skill | With skill |
|---|---|---|
| States that 1 to 3 isolated rules result in a 70–89 score while more than 3 drop the score below 70. | Missed | Passed |
The instructions your model reads when the skill loads.
对当前平台配置进行 7 维度评分，定位能力短板并给出优先改进建议。
/harness-audit
或由 harness-optimizer specialist 直接调用。
检查所有角色是否有对应的 agents/roles/*.md 生成文件，specialist agents 是否与 commands/ 命令一一对应。
评分标准：
检查 skills/、skills/、skills/ 中所有技能是否有有效 SKILL.md，role.yaml 中引用的技能是否实际存在。
评分标准：
检查 hooks/hooks.json 配置是否启用关键 hook，session_start.py/session_end.py 是否能实际运行。
test cases passed, vs 4 / 19 without
More skills from the same source and nearby in the registry, with their headline benchmark lift.