一键导入
skill
mokata · Author a new mokata skill test-first (RED-GREEN-for-docs); the write is human-gated.
用 Codex 或 Claude 帮你安装 复制这段 Prompt,粘贴到 Codex、Claude 或其他助手里,让它检查 Skill 页面并帮你完成安装。
菜单
mokata · Author a new mokata skill test-first (RED-GREEN-for-docs); the write is human-gated.
用 Codex 或 Claude 帮你安装 复制这段 Prompt,粘贴到 Codex、Claude 或其他助手里,让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
mokata · Socratic pre-spec exploration — explore approaches WITH the user, one question at a time, and HARD-GATE the spec behind explicit approval. Runs standalone or as the front of mokata's pipeline.
mokata · Implement the minimum to turn a failing test green.
mokata · Turn the problem into testable acceptance criteria; map each to a test.
mokata · API & interface design — contract-first, blast-radius on every change.
mokata · Browser testing (DevTools) — live runtime captured via the existing MCP surface, fed to the test gate.
mokata · Start from a reproducer and a failing test, then fix.
| name | skill |
| description | mokata · Author a new mokata skill test-first (RED-GREEN-for-docs); the write is human-gated. |
| argument-hint | <name> --summary <s> --require <doc>:<must-contain> --content-file <f> |
| allowed-tools | Bash, Read, Write |
Author a new /<name> skill the mokata way: test-first for docs (G6). You declare what
the skill's content MUST contain (--require name:must-contain, repeatable); the draft is
checked against those requirements (RED) and only a passing draft (GREEN) is promoted
to a command. Writing the rendered command is a durable write, so it is human-gated.
Draft the content to a markdown file (use Write), then decide the doc requirements it
must satisfy — each as --require <name>:<must-contain>.
Resolve the bundled engine (read ~/.mokata/plugin-root → ROOT, or a mokata CLI on
PATH):
PY="$(command -v python3 || command -v python)"
ENGINE="PYTHONPATH=\"$ROOT/src\" \"$PY\" -m mokata"
Check it RED→GREEN first (writes nothing while RED):
eval "$ENGINE skill author <name> --summary \"mokata · <one-liner>\" \
--require <doc>:<must-contain> --content-file <draft.md> --path ."
If it reports RED (a requirement unmet), revise the draft until every requirement passes.
Approve the write — only after the draft is GREEN and the user confirms, add --yes:
eval "$ENGINE skill author <name> --summary \"mokata · <one-liner>\" \
--require <doc>:<must-contain> --content-file <draft.md> --yes --path ."
The write goes through the universal gate (a secret in the content is hard-blocked even when approved). Report where the new skill was written.