instructions-authoring
A plugin that ships instruction-authoring skills for AI agents. It covers cross-model authoring, controlled-English instruction writing, readiness assessment, and detailed instruction diagnostics.
What ships
skills/cross-model-instruction-authoring/— create or revise Agent Skills, custom agent prompts, subagent instructions, and other instruction packages for use across selected models and runtimes, separating portable behavior from runtime-specific adapters.skills/agent-skill-audit/— assess whether an agent instruction or skill package is ready for its target models and runtime, with a focus on discovery, completeness, portability, and maintainability.skills/instruction-quality-audit/— diagnose exact defects in AI instruction artifacts, including contradictions, ambiguity, authority conflicts, closure gaps, duplicated rules, and output-contract problems.skills/controlled-agent-instruction-english/— write, rewrite, or audit model-facing operational prose for consistent terminology, explicit modality and scope, conditions, precedence, and failure behavior.
Skill names
Use these names to select a skill where the host supports explicit invocation:
cross-model-instruction-authoring
agent-skill-audit
instruction-quality-audit
controlled-agent-instruction-english
Where each skill fits
| Stage | Skill |
|---|---|
| Designing a reusable instruction package | cross-model-instruction-authoring |
| Auditing a skill package for readiness | agent-skill-audit |
| Finding exact instruction defects | instruction-quality-audit |
| Writing or auditing precise operational prose | controlled-agent-instruction-english |
Scope
- Skill instructions and references only. No hooks.
- Skill descriptions state when they apply. Actual discovery and explicit invocation syntax depend on the host.