소스 정보
- 저장소
- imbue-ai/catalyst
- 최근 소스 활동
- 2026년 5월 28일 00:12
- 감지된 SKILL.md 언어
- 영어
- 스타
- 32
- 포크
- 2
설치 방법
기본적으로 소스를 먼저 확인하는 Prompt가 선택됩니다. 직접 명령으로 전환하거나 로컬 사본을 다운로드할 수도 있습니다.
소스 파일 검토
설치 여부를 결정하기 전에 SKILL.md와 SkillsMP에 표시된 보조 파일을 읽어 보세요.
메뉴
기본적으로 소스를 먼저 확인하는 Prompt가 선택됩니다. 직접 명령으로 전환하거나 로컬 사본을 다운로드할 수도 있습니다.
설치 여부를 결정하기 전에 SKILL.md와 SkillsMP에 표시된 보조 파일을 읽어 보세요.
Codex 또는 Claude로 설치 이 Prompt를 복사해 Codex, Claude 또는 다른 어시스턴트에 붙여 넣으면 Skill 페이지를 검토하고 설치를 진행할 수 있습니다.
직접 명령은 검토 Prompt를 거치지 않습니다. 실행하기 전에 소스를 확인하세요.
npx skills add https://github.com/imbue-ai/catalyst --skill write-different-theory명령은 한 줄로 유지됩니다. 복사하기 전에 가로로 스크롤해 전체 내용을 확인하세요.
로컬 사본을 원하시나요? SkillsMP에서 현재 제공할 수 있는 파일을 다운로드하세요.
SKILL.md 표시 중
SOC 직업 분류 기준
| name | write-different-theory |
| description | Write a theory taking a *different* approach from the provided ones. |
| argument-hint | A list of prior theory IDs (T_20260414_..., ...), optional literature ID (e.g. L_20260414_...). |
You are an expert scientific agent. Your goal is to develop a theory to explain a given phenomenon. Some previous attempts have been made to explain the phenomenon, but they turned out to be unsatisfactory. So your goal is to AVOID taking the same approaches as the previous theories, and instead come up with a different explanation that accounts for the phenomenon and is supported by evidence.
Arguments: $ARGUMENTS
The arguments contain a list of prior theory IDs (like T_20260414_...), and an optional literature review ID (like L_20260414_...). Parse all IDs from the arguments.
All commands must be run in the current working directory. Do not cd anywhere else, do not try to use the global /tmp folder or TMPDIR (only use the local ./tmp folder).
Set up a folder to hold the pre-existing theories:
CONTEXT_DIR: mktemp -d -p ./tmp write-different-theory-context-XXXX
Run this command to populate the context:
uv run python <SKILL_BASE_DIR>/scripts/context_manager.py create_context --for_agent_type write-different-theory --target_folder <CONTEXT_DIR> --from_theory <THEORY_ID_1> [--from_theory <THEORY_ID_2> ...]
<CONTEXT_DIR>/theories/<theory_id>/theory.md — the previous theory attempts that we're trying to be different from.context_manager.py.phenomenon.txt in the current work directory. If so, read the phenomenon description from there. Otherwise, if the phenomenon.txt file does not exist, read the first section(s) of one of the theory files. You can pick any of the theories for this step, as they should all be targeting the same phenomenon.<CONTEXT_DIR>/theories/<theory_id>/theory.md. Understand the approaches they took, identify the assumptions they made, the angles they explored, and the evidence they used. This will help you figure out how to take a different approach.write-theory skill, passing to it the phenomenon description, and the literature review ID if available. Follow all instructions and execution steps provided in that skill, BUT make sure that the new theory takes a FUNDAMENTALLY DIFFERENT approach from all of the previous theories. You might have to iterate and explore for a while to find a different angle, so please take your time and be creative. It is expected that you have to run through multiple cycles of hypothesis development and experimentation before you arrive at a properly novel perspective. When the write-theory skill is done, it will provide a new theory ID (e.g. T_20260414_143100_d4e5f6) that forms the result of this skill.