メニュー
Stream-JSON chaining for sequential multi-agent pipelines and data transformation. Use when step N's output must feed step N+1's input (build->test->optimise, extract->transform->validate). Not for parallel swarms or GitHub CI.
デフォルトでは、最初にソースを確認する Prompt が選択されています。直接コマンドに切り替えるか、ローカルコピーをダウンロードすることもできます。
インストールを決める前に、SKILL.md と SkillsMP に表示されている付属ファイルをお読みください。
SOC 職業分類に基づく
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。
直接コマンドでは確認用 Prompt が省略されます。実行前にソースを確認してください。
npx skills add https://github.com/DreamLab-AI/agentbox --skill stream-chainコマンドは1行のまま表示されます。コピー前に横へスクロールして全体を確認してください。
ローカルで確認しますか？SkillsMP が現在取得できるファイルをダウンロードできます。
SKILL.md を表示中
Implement features with tests and quality gates, debug hard bugs, and stress-test designs. Use when building a feature with TDD/EDD, chasing a stubborn multi-function bug (feedback-loop-first), interrogating a design before coding, or running a coverage/security/accessibility quality-gate pass. Supersedes agentic-qe, reasoningbank-*, and pair-programming.
Plan-execute-reflect loop over a persistent Python kernel (the `code-interpreter` MCP): write Python, execute, observe the returned trace (stdout / exception / last value), then revise. Variables, imports, and dataframes survive across every tool call in the session. Use for multi-step numerical reasoning, data wrangling, scientific Q&A, hypothesis-test loops, or any task where intermediate state must persist between turns and a runtime result should drive the next step.
Fires as a post-task hook after a non-trivial completed task (3+ tool calls, observable terminal outcome) to distil 0-N generalisable IF/THEN lessons from the trajectory and store them in RuVector so they surface at the start of similar future tasks; also invocable manually to record a lesson after a complex or failed task. NOT for trivial (<3 tool calls) or interrupted tasks, and NOT a general memory-write tool — it writes only ex:DistilledLesson records to the code-harness-lessons namespace. Schema, gates, prompt, and contradiction/archive policy live in references/.
| name | stream-chain |
| description | Stream-JSON chaining for sequential multi-agent pipelines and data transformation. Use when step N's output must feed step N+1's input (build->test->optimise, extract->transform->validate). Not for parallel swarms or GitHub CI. |
| version | 1.0.0 |
| category | workflow |
| tags | ["streaming","pipeline","chaining","multi-agent","workflow"] |
Execute multi-step workflows where each agent's complete output flows into the next step as context, enabling sequential data transformation and multi-agent coordination.
Two modes:
run) — your own prompt sequence, full control.pipeline) — battle-tested workflows for common tasks.swarm-advanced (this skill is sequential).github-workflow-automation.flow-nexus-swarm.Custom chain (minimum 2 prompts; output flows step-to-step):
claude-flow stream-chain run \
"Analyze codebase structure" \
"Identify improvement areas" \
"Generate action plan"
Predefined pipeline:
claude-flow stream-chain pipeline analysis # or: refactor | test | optimize
Full QA sweep (analysis -> refactor -> test -> optimize) via helper script:
./scripts/qa-chain.sh 60 # arg = per-step timeout seconds
Common options for both modes: --verbose, --timeout <seconds> (default 30),
--debug.
.claude-flow/config.json.