Designs whether and how to create, merge, split, or rescope Agent Skills before implementation.
npx skills add https://github.com/mtk177a/skills --skill design-skill-mtk177a
Please help me install this Agent Skill. Skill: design-skill Source: https://github.com/mtk177a/skills/tree/main/skills/design-skill Command: npx skills add https://github.com/mtk177a/skills --skill design-skill-mtk177a
name and description metadata with negative boundaries to prevent collisions. - Evaluation-First Planning: Maps each material claim to a plausible failure, scenario, and grading method before extensive authoring, and produces an implementation-ready handoff without editing files. - Use Case: When two installed Skills both activate for the same task, use this Skill to decide whether to merge them, keep them split, or rescope, including coexistence risks and evaluation cases. ## Quick Start Ask the agent to decide whether a proposed new Skill is the right intervention and to produce a design with responsibility, trigger boundary, and evaluation plan before any implementation.High-intent search queries and answers about installing and using this skill.
Compare the relevant interventions first: no durable guidance, a local instruction or tool, updating or merging an existing Skill, splitting an overloaded one, or creating a new Skill. Prefer no Skill when the model already performs the task reliably without reusable specialized context.
Prefer updating or merging when the trigger, output contract, and risk boundary remain coherent. Prefer a new or split Skill when combining responsibilities would make triggering, loading, or execution ambiguous.
Put the complete implicit-trigger context in the description, including what the Skill does and when it should trigger, front-loading key terms. Include a negative boundary when it prevents a likely collision, since clients select Skills from metadata before loading the body.
No. The workflow stops at an implementation-ready handoff covering the selected intervention, evidence, metadata draft, resource allocation, safety properties, and evaluation plan. Editing files, adding dependencies, or changing repository policy is left to an authorized implementation workflow.
Map each material responsibility, trigger boundary, and changed behavior to a plausible failure, a scenario or check, and a grading method. Include only dimensions addressing distinct risks, and state whether static validation, targeted regression, or empirical tuning is warranted.
Record the upstream source, pinned revision, license, complete file set, and local modifications. Trace the full data and capability path from input through scripts and tools to outbound destinations, and reject or leave the design unverified when provenance or data flow cannot be established.
dotnetWrite structured postmortems for production regressions with root cause and prevention analysis.
Shockwave2kDraft SKILL.md files and run evaluation loops to measure trigger accuracy.
Rosetears520Standardize OpenCode Agent Skill creation with SKILL.md structure and validation checks.
ayehiaaInitialize a Git repository and create an initial commit.