agentleFS
BaratiLabBaratiLab/mechanical-engineering-skills/skills/sls-process-optimization/SKILL.md

sls-process-optimization

Skill16 starsChanged 16 days agoMITView on GitHub

Optimize polymer selective laser sintering (SLS) process parameters batch by batch — keep a build history, reduce ASTM D638 tensile and D790 flexural test data, compare against the material datasheet, and propose the next batch's fill energy density, hotspot overlap and surface temperature with a bounded Bayesian suggester, the agent explaining each change. Based on the Barati Farimani Lab's agentic SLS work (PA12 GF, PA11, PA12 blends on an SLS4All Inova). Use for SLS parameter development, new powders, or learning from ASTM results.

The file

We split this skill into the rules it gives an agent

5 rules. About 0 are common, 2 are distinctive, and 3 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈2

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 3

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Add the user's powder with its datasheet targets and DSC melt peak: material --name "PA12 White" --tds Et=1950,UTS=47 --tmelt 186.

    line 18
  2. Very specific:

    First batch of a new powder: Follow the machine's patch test (the paper prints a 5 × 5 grid of energy settings) and start from the lowest …

    line 20
  3. Very specific:

    quote the scatter: If one batch can't be reproduced, mark it --exclude.

    line 32
  4. Distinctive:

    Use the notes (warping, brittle corners, early build stop) to override the model when needed.

    line 40
  5. Distinctive:

    Every proposal needs a human go-ahead before printing (the paper's GUI step).

    line 48

≈ “Common” is an estimate: at least 1 in 5 of the 372,429 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.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

SLS process optimization (agentic)

Pak, Alvarado & Barati Farimani (arXiv:2608.25928, 2026) let an agent remember every build and its ASTM results, then propose the next batch on an open-source SLS printer. The agent reached datasheet values for UTS and flexural strength in 3–4 batches per material. scripts/sls.py is that memory, the test-data reduction, and a bounded suggestion engine. You reason, explain, and let the user approve. Paper data and numbers are in references/paper-notes.md.

Workflow

  1. Set up the history.
    • python scripts/sls.py --db builds.sqlite init --seed-paper loads the paper's 21 batches (PA12 GF, PA11 Onyx, PA12 Blend). Omit --seed-paper for an empty history.
    • Add the user's powder with its datasheet targets and DSC melt peak: material --name "PA12 White" --tds Et=1950,UTS=47 --tmelt 186.
  2. First batch of a new powder. Follow the machine's patch test (the paper prints a 5 × 5 grid of energy settings) and start from the lowest energy that sinters visibly. If nothing has been recorded, suggest returns a conservative starting point.
  3. After each build, reduce the test data.
    python scripts/sls.py astm --test d638 --type 4 --csv t1.csv t2.csv t3.csv t4.csv t5.csv > tensile.json
    python scripts/sls.py astm --test d790 --csv f1.csv f2.csv f3.csv > flex.json
    python scripts/sls.py --db builds.sqlite result --batch Q --from tensile.json [--Ef .. --sf ..]
    • Each CSV needs a load (or force) column in N and an extension/deflection column in mm. Pass --strain if the second column is already strain.
    • The output includes mean ± sd and the best specimen. Record the best (paper convention), but quote the scatter. If one batch can't be reproduced, mark it --exclude.
  4. Propose the next batch. python scripts/sls.py --db builds.sqlite suggest --material "PA11 Onyx" [--objective UTS|Et|score]
    • Candidates come ranked by expected improvement, within machine and step bounds, and list which levers changed.
    • Pick one, and change one or two levers. Explain the physics:
      • more fill energy and overlap → better coalescence, until warping or degradation sets in;
      • higher surface temperature → less curl and better layer bonding, until the powder cakes.
    • Use the notes (warping, brittle corners, early build stop) to override the model when needed.
    • batch refuses a surface temperature within 3 °C of the melt peak.
  5. Track progress: history --material ... --plot progress.png shows every batch against the datasheet targets.

Rules

  • The datasheet is the target, but machine and atmosphere matter. PA11 Onyx is specified for nitrogen, which the paper's machine lacked. Say when a target is unreachable on this setup.
  • Every proposal needs a human go-ahead before printing (the paper's GUI step). Never push parameters to a printer without approval.

Cite: Pak, Alvarado & Barati Farimani, arXiv:2608.25928 (2026).

Download now (ZIP)

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
What did you use it for? e.g. Reviewing pull requests in a Rails app

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.