Install any skill in seconds. Free to start, no credit card required.
Get Started Free →shinpr/llm-friendly-context
Clarifies inputs, outputs, success criteria, decisions, and unresolved conditions so downstream agents can execute without guessing.
Passed 22 of 23 test cases with the skill, and 14 of 23 without it.
and 8 more in the full report
8 fixed14 passed either way1 still failing0 broke
Tokens, turns and time describe resource use, not a dollar price.
One of 8 test cases the skill fixed. Same model, same prompt, with and without the skill.
我们正在为 Database Agent 制定安全与规范原则。团队成员写了一句:“严禁执行 DROP TABLE 指令。”请重写此禁令,使其符合核心规则中关于保护不可逆边界的要求。
| Checked | Without skill | With skill |
|---|---|---|
| Preserves the prohibition on destructive commands while stating explicit allowed schema migration actions | Missed | Passed |
The instructions your model reads when the skill loads.
目标是实现稳定的下游执行:下一个智能体应当知道要读什么、要做什么、什么算成功,以及哪些未解决的决策会改变结果。
本技能规范面向 LLM 的输出,使提示词、交接和生成的产物清晰明确。调用方提供产物类型以及任何产物特定的模板或输入契约;只包含其使用方用来决策、行动或验证所需的信息。当使用方根据字段分支处理时,使用已声明契约的字段名和取值含义。
使用正向、可执行的指令
将模糊指令具体化
appropriate(合适)、proper(恰当)、related(相关)、existing behavior(现有行为)、optional(可选)、as needed(按需)、if needed(如有需要)、per convention(按惯例)、未解决的备选方案、TBD(待定)、placeholder(占位符)明确输出形态
提供最小且充分的上下文
将复杂工作拆解为可验证的步骤
明确允许不确定性
保持约束的适度性
test cases passed, vs 14 / 23 without
More skills from the same source and nearby in the registry, with their headline benchmark lift.