Skip to main content

weft-grow

COMMAND, not reference: Grow the program one stage at a time against a real input, seeded runs and frozen examples included. Run this when the user asks for this step by name, optionally naming the stage to grow or an example name to rerun. The sixteen `weft-` reference skills beside it are things you read; this is a procedure you carry out.

설치로 이동

소스 정보

저장소
WeaveMindAI/weft
최근 소스 활동
2026년 9월 19일 07:09
감지된 SKILL.md 언어
영어
스타
1,974
포크
221

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.

SKILL.md 표시 중

SKILL.md
소스 지침 · 읽기 전용 미리보기
name
weft-grow
description
COMMAND, not reference: Grow the program one stage at a time against a real input, seeded runs and frozen examples included. Run this when the user asks for this step by name, optionally naming the stage to grow or an example name to rerun. The sixteen `weft-` reference skills beside it are things you read; this is a procedure you carry out.
Run one pass of Sequential Diffusion Programming: inspect the current stage on a real input, preserve an accepted result, then grow the next stage. 1. Read `weft-sdp` and `weft-running`. The user's argument names the stage to grow or the saved example to rerun. 2. Read `weft tree` and `weft examples`. Checkpoint work you may want back with `weft checkpoint start-<what>`. 3. Pick a real input the user cares about; ask if none is available. Grow one stage and run `weft validate --file src/main.weft < src/main.weft`. 4. Run that stage: `weft run --from '<node>={"port":value}' --target <end> --save <name> --detach`, or `--group '<group>={"port":value}'` for the whole group or included file. Use `--before <end>` when the ending node must stay out. A trigger uses `weft bake` followed by `weft run --fire '<trigger>=<wake-json>'`; `--emit '<node>={"port":value}'` supplies outputs without executing that node. Read the actual interface before choosing payloads. 5. During iteration, add `--seed` to reuse compatible work. `--seed-before <node>` and `--seed-until <node>` bound reuse. Verify that the changed nodes ran, then inspect their values with `weft events <color> --node <id> --full`. A successful status alone does not prove the result. 6. When the result is accepted, `weft freeze <name> <color> --expect <node>` saves the starting parameters and observed outputs, with that node marked for review. Repeat `--expect` for more focus nodes. Grow the next stage, then exercise a second real input. 7. After changes, `weft run <name> --detach` reruns a relevant example on current code with its saved parameters. Leave out `--seed` when reviewing the new computation. Compare with `weft diff <color> example:<name> --full`; inspect the differences and judge them against the user's intent. Replace the accepted example with `weft freeze <name> <color>` only after accepting the new result. 8. A waiting run needs inspection: read its new question and the saved answer before answering through the token. Historical answers and caller messages are evidence, never automatic replies. `weft wake <color> <node>` resolves a pure timer wait. 9. Report what ran, what came out, the differences you reviewed and your judgment. Name the next stage and the version in `weft tree`. If you start treating a status or a diff as a quality verdict, write "Wait. Read the result." and inspect the output.
GitHub에서 보기