pipeline-builder
Skill114 starsChanged 5 months agoMITView on GitHub
Break a big, multi-part task into a clean pipeline where each step's output feeds the next, with context carried forward. Use when one prompt is trying to do too much and the result drowns. Works for any multi-stage task in any field.
1 other project uses a version of this.
Adopted by
Most recently taken up by
- tezansahu/ai-garagetheir own version, first seen Aug 2, 2026
Other versions1 project uses a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 1
- tezansahu/ai-garage★ 16same name
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository6 other files it gives its agents.
- connector-setupSkill
- context-brainSkill
- orchestratorSkill
- prompt-architectSkill
- skill-authorSkill
- work-engineSkill
The file
We split this skill into the rules it gives an agent
3 rules. About 0 are common, 3 are distinctive, and 0 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈3
Distinctive · estimate
This author's own choices. Worth a look.
0
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 20
Write it out as ordered stages.
- Distinctive: line 29
Keep stages single-purpose.
- Distinctive: line 42
Always carry the brief/context forward — drift between stages is the main failure mode.
≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Pipeline Builder
Big asks fail when research, making, and adapting get crammed into one prompt. The fix is a pipeline: each step does one job, and its output becomes the next step's input. This skill designs that pipeline with the user.
Role-agnostic. "Research → draft → repurpose" is one example shape; a hiring loop, a design review, or a financial close are others. Ask.
Step 1 — Interview
- What's the big task? The whole thing they're trying to do in one go.
- What are the natural stages? Help them name 2–4 steps where each produces something the next one needs. If they can't, propose stages from the task and confirm.
- What context must carry forward? The facts, brief, or source-of-truth that every stage has to respect so nothing drifts.
Step 2 — Design the pipeline
Write it out as ordered stages. For each stage:
STAGE n — <name>
Input: <what it receives (incl. carried context)>
Does: <the one job>
Output: <the artifact it hands to the next stage>Keep stages single-purpose. The first stage's output is usually not a draft — it's the brief/analysis the making stage needs. Carry the original brief into every stage so the work stays on-message.
Step 3 — Run it as steps, not one prompt
Have the user run each stage as its own prompt (or sub-agent), passing the prior output + carried context forward. Check the hand-off between stages — that's where drift hides.
Step 4 — Save the shape
Once it works, the stage list is a reusable template. Suggest packaging it with skill-author so the whole pipeline runs from one trigger next time, and feeding any per-stage prompts through prompt-architect.
Rules
- One job per stage. If a stage does two things, split it.
- Always carry the brief/context forward — drift between stages is the main failure mode.
- Wall-clock and quality both improve when stages are clean; don't collapse them back into one prompt.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

