メニュー
Runs checks from .ged/work/<work-id>/TESTS.md after task implementation, summarizes pass/fail, and prepares retry briefs. Triggers include "verify", "test", "check", "did it work", or after completing an implementation task.
Codex または Claude でインストール この Prompt をコピーして Codex、Claude、または他のアシスタントに貼り付けると、Skill ページを確認してインストールできます。
SOC 職業分類に基づく
Discover external skills during GedPi's skill-fit checkpoint. When the current bundled, project, and user skills don't cover a task's needs, search the open ecosystem via npx skills, evaluate candidates, and install through GedPi's project-skill mechanism. Triggers include "how do I do X", "find a skill for X", "is there a skill for X", "skill for", "need a skill", "search for skills", "skill-fit", "ecosystem"
Implements the next task from .ged/work/<work-id>/TASKS.md with narrow scope and concise handoff notes. Triggers include "do the next task", "implement", "execute", or when a task is ready for implementation.
Produces or refines implementation plans in .ged/work/<work-id>/SPEC.md, task breakdowns in .ged/work/<work-id>/TASKS.md, and verification criteria in .ged/work/<work-id>/TESTS.md. Triggers include "plan", "break down", "what tasks", "how to implement", or when starting a new feature.
Use before planning or implementing a non-trivial change when the request is not fully clear. Clarifies goals, desired behavior, constraints, edge cases, non-goals, tests, rollout, and success criteria by asking one concise question at a time with a recommended answer/default.
Grilling session that challenges your plan against the existing domain model, sharpens terminology, and updates documentation (CONTEXT.md, ADRs) inline as decisions crystallise. Use when user wants to stress-test a plan against their project's language and documented decisions. Triggers include "domain", "glossary", "terminology", "ubiquitous language", "CONTEXT.md", "ADR", "architecture decision", "decision record", "context map", "domain model".
Create a project-local skill when the skill-fit checkpoint finds no adequate existing skill. Use when find-skills returns nothing suitable, the task gap is reusable at the project level, and you need to capture domain knowledge, workflow patterns, or project-specific procedures as a SKILL.md. Triggers include "create a skill", "new skill", "write a skill", "make a skill for", "capture this", "turn this into a skill", "project skill", "custom skill"
| name | ged-verification |
| description | Runs checks from .ged/work/<work-id>/TESTS.md after task implementation, summarizes pass/fail, and prepares retry briefs. Triggers include "verify", "test", "check", "did it work", or after completing an implementation task. |
.ged/work/<work-id>/TESTS.mdFor non-trivial Ged tasks, planning should happen after classification, any needed grill-me clarification, and skill-fit resolution. With subagents enabled, ged-explorer performs read-only skill inventory/evaluation/search and reports findings; the main brain then installs project-scoped external skills or creates narrow project-local skills when warranted. In solo/no-subagent mode, the main brain performs the skill-fit checkpoint itself. Do not install global/user skills automatically.