From build-pipeline
Run a dynamic workflow on a real input, observe it without flooding context, verify its assumptions (gates fire, structured returns thread and branch, parallel stages overlap), then turn what you learn into edits to the workflow, its agents and skills, and the build-pipeline generator that produced it. Use after generating or changing a pipeline, or to dogfood and harden any Claude Code dynamic workflow.
How this skill is triggered — by the user, by Claude, or both
Slash command
/build-pipeline:improve-pipeline <workflow-name> <concrete input to run it on><workflow-name> <concrete input to run it on>opusThis skill is limited to the following tools:
The summary Claude sees in its skill listing — used to decide when to auto-load this skill
Work in the main context; do not fork. This loop drives the Workflow tool, a Monitor, and task notifications, which only the main loop can orchestrate (a subagent cannot spawn subagents).
Work in the main context; do not fork. This loop drives the Workflow tool, a Monitor, and task notifications, which only the main loop can orchestrate (a subagent cannot spawn subagents).
Input: $ARGUMENTS is the workflow name followed by a concrete input to run it on.
npm run dev for a stage that drives the browser) and confirm it is up before launching.args. Keep the transcript dir and runId from the result.agentType, reload first. A mid-session agent file does not register: agent({agentType}) resolves against the registry frozen at session start.git status --short and emit when the changed-file set changes.<transcript-dir>/journal.jsonl. It maps each agentId to its started and result records and holds each stage's return value, so it is the spine of the run.agent-*.jsonl. Never read one whole: a subagent transcript is the full conversation and overflows context.gate not satisfied. A passing gate is silent, so absence proves nothing. To prove a gate fires, catch a stage that stops while its gate is unsatisfied, or run a throwaway probe agent whose gate appends to a log file (always exit 0 so it cannot loop), and reload so its agentType resolves.started before either shows result.For anything that did not progress as intended, trace it to the exact stage, agent, or workflow line, reading that one transcript with targeted greps.
Guides creation, editing, and verification of skills for AI coding agents using test-driven development with subagent scenarios. Use when authoring or debugging skills.
npx claudepluginhub hackersandwizards/agentic-engineering-training-nextjs --plugin build-pipeline