Skip to content
cccskills
FreePublished on GitHub

run-eval

Run the LangSmith evaluation suite and display pass/fail results

The English summary is being prepared. Showing the original description.

See how to install

Files included (1)

  • SKILL.md1.8 KB

SKILL.md (original)

Review the instructions the agent will receive before you install.

/run-eval

Steps

  1. Ensure Docker is running

    docker compose ps
    

    If PostgreSQL is not up:

    docker compose up -d
    
  2. Run the evaluation suite

    uv run dotenv -f .env run -- python scripts/run_eval.py
    

    Note the experiment name printed (e.g. medox-<hash>).

  3. Fetch and display results Write the following script to /tmp/check_eval.py then run it:

    from langsmith import Client
    
    client = Client()
    runs = list(client.list_runs(project_name='<experiment_name>', is_root=True))
    print(f'Eval cases: {len(runs)}')
    print()
    
    passed, failed = 0, 0
    for run in runs:
        fb = list(client.list_feedback(run_ids=[str(run.id)]))
        score = fb[0].score if fb else None
        comment = fb[0].comment if fb else ''
        prompt = (run.inputs or {}).get('prompt', '').strip()[:75]
        status = 'PASS' if score == 1 else 'FAIL'
        if score == 1:
            passed += 1
        else:
            failed += 1
        print(f'[{status}] {prompt}')
        if comment and comment != 'OK':
            print(f'       -> {comment}')
    
    print()
    print(f'Result: {passed} passed, {failed} failed out of {len(runs)}')
    

    Replace <experiment_name> with the value printed in step 2, then:

    uv run dotenv -f .env run -- python3 /tmp/check_eval.py
    
  4. Investigate failures For any [FAIL], read the comment and:

    • Check the relevant node/tool in src/medox/agent/
    • Check the evaluator logic in scripts/run_eval.py
    • Use /add-eval-case to add a regression case if a new edge case was found
  5. Report summary Print the final Result: N passed, M failed out of X line to the user.

Reviews

No reviews yet. Tell others what you thought.

Comments

No login needed. Share how it worked for you, ask a question, or note where it did not run.

No comments yet. Be the first to write one.

Comments are published right away. Do not include personal information such as email addresses or phone numbers. Up to 2 URLs.

0 / 1,000 characters

This form is protected by reCAPTCHA, and the Google Privacy Policy and Terms of Service apply.

More skills from this repository

Add a tool function to the Medox agent. Use when creating a tool file, registering it in the API loop, and writing its test.

The English summary is being prepared. Showing the original description.

spideystreet/medox13Updated Oct 1, 2026

Add a new E2E test case to tests/e2e/prompts.yaml for LangSmith evaluation. Use when adding interaction pairs to test, covering new ANSM drug classes, or rebalancing eval coverage.

The English summary is being prepared. Showing the original description.

spideystreet/medox13Updated Oct 1, 2026

commit

Free

Create a git commit with Medox conventions (co-author, conventional type/scope). Invoke manually with /commit — do NOT trigger automatically.

The English summary is being prepared. Showing the original description.

spideystreet/medox13Updated Oct 1, 2026

Audit code changes on the current branch and update docs/ + .claude/rules/ to stay in sync. Run before /pr.

The English summary is being prepared. Showing the original description.

spideystreet/medox13Updated Oct 1, 2026

lint

Free

Run all project linters (ruff, mypy, sqlfluff) and report results. Invoke with /lint.

The English summary is being prepared. Showing the original description.

spideystreet/medox13Updated Oct 1, 2026

Add a new Dagster asset to the pipeline. Use when creating a Bronze, Silver, or Gold layer asset.

The English summary is being prepared. Showing the original description.

spideystreet/medox13Updated Oct 1, 2026

See all skills by spideystreet

Report a problem with this skill