agentsclimarketplace

Submission evaluator

Skill alainlebret/claude-agents/higher-ed-teaching-agents/skills/submission-evaluator

Modular skill and workflow ecosystems for AI coding assistants. Structured, multi-agent pipelines for specialized professional domains.

Install
npx -y skills add alainlebret/claude-agents --skill submission-evaluator

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

One thing to look at

  • 0 stars0 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

What its author says it does

Copied from the file, not written here

Evaluates student submissions for practical assignments, machine exams, and projects using an evidence-first workflow. Use when an instructor needs robust grading from a subject, rubric, reference materials, tests, and student files, especially in programming and systems courses.

SKILL.md

1.2 KB, as published. Nobody here has run it

Submission Evaluator

Goal

Grade submissions through structured evidence, not impressionistic judgment.

Mandatory workflow

  1. Read the subject.
  2. Read the rubric.
  3. Read the reference materials.
  4. Collect observable evidence.
  5. Compute criterion scores.
  6. Generate the final report.

Required artifacts

When grading many submissions, create:

  • evidence.json or per-submission evidence files;
  • a summary of criterion scores;
  • a final report in Markdown.

Evidence policy

Record at least:

  • expected files;
  • present files;
  • build status;
  • test results;
  • interface conformity;
  • major defects;
  • strengths;
  • awarded points per criterion.

Rules

  • Never assign final grades before evidence collection.
  • If build fails, analyze sources anyway.
  • Give partial credit only when supported by evidence.
  • Keep student identity distinct from path anomalies or symlink cases.

Keep looking

Skills are one crate of 328,083. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.