improve-context
The harness concierge for improving a project's context — the human-in-the-loop expert on every context lever (the codebase's own shape, AGENTS.md, the Expert long-term memory, /intent, local-check lints) and on STUCK forensics. Use to resolve a STUCK PR (diagnosis-first), to improve any lever ("improve long-term memory with X", "our folder structure is a mess"), or with no args to present the lever map — what each one is and when it's worth pulling — and let the human choose. Drives horizontal refactors that make the codebase legible to the next agent, and builds evals over the project's own context (a per-lever pyramid under evals/, graded on a developer-intent rubric). Replaces /evaluate-sessions. Triggers - improve-context, improve context, unstick, diagnose stuck, STUCK, evaluate sessions, review the build trail, improve long-term memory, seed the Expert, tune AGENTS.md, lint quality, harness evals, agent legibility, harnessability, folder structure, naming, our code is a mess, refactor for the agent.
Informações da origem
- Repositório
- capitalone/context-specs
- Última atividade na origem
- 22 de julho de 2026 às 17:15
- Idioma detectado do SKILL.md
- inglês
- Estrelas
- 45
- Forks
- 6
Opções de instalação
Por padrão, está selecionado o prompt que primeiro revisa a origem. Você pode mudar para um comando direto ou baixar uma cópia local.
Revise os arquivos de origem
Leia o SKILL.md e os arquivos complementares exibidos pelo SkillsMP antes de decidir se vai instalar.
Exibindo SKILL.md
| name | improve-context |
|---|---|
| description | The harness concierge for improving a project's context — the human-in-the-loop expert on every context lever (the codebase's own shape, AGENTS.md, the Expert long-term memory, /intent, local-check lints) and on STUCK forensics. Use to resolve a STUCK PR (diagnosis-first), to improve any lever ("improve long-term memory with X", "our folder structure is a mess"), or with no args to present the lever map — what each one is and when it's worth pulling — and let the human choose. Drives horizontal refactors that make the codebase legible to the next agent, and builds evals over the project's own context (a per-lever pyramid under evals/, graded on a developer-intent rubric). Replaces /evaluate-sessions. Triggers - improve-context, improve context, unstick, diagnose stuck, STUCK, evaluate sessions, review the build trail, improve long-term memory, seed the Expert, tune AGENTS.md, lint quality, harness evals, agent legibility, harnessability, folder structure, naming, our code is a mess, refactor for the agent. |