Skip to content

feat: improve launching-evals and nel-assistant skills#910

Open
piojanu wants to merge 5 commits intomainfrom
pjanuszewski/enforce-post-run-monitoring
Open

feat: improve launching-evals and nel-assistant skills#910
piojanu wants to merge 5 commits intomainfrom
pjanuszewski/enforce-post-run-monitoring

Conversation

@piojanu
Copy link
Copy Markdown
Contributor

@piojanu piojanu commented Apr 14, 2026

Summary

  • Makes post-run monitoring a mandatory step after every nel run by updating the skill workflow to poll until terminal state
  • Clarifies where to find the invocation ID in nel run output
  • Adds resume command documentation to launching-evals skill
  • Adds multi-node switches: 1 performance tip to nel-assistant skill
  • Adds multi-instance scale-down example for quick validation runs to nel-assistant skill
  • Uses tool-agnostic polling instruction (not Claude Code-specific /loop)

Test plan

  • Verify skill triggers monitoring after nel run
  • Verify resume command appears in launching-evals skill reference
  • Verify multi-node and scale-down tips render correctly in nel-assistant skill

🤖 Generated with Claude Code

@copy-pr-bot
Copy link
Copy Markdown

copy-pr-bot bot commented Apr 14, 2026

Auto-sync is disabled for draft pull requests in this repository. Workflows must be run manually.

Contributors can view more details about this message here.

@piojanu piojanu marked this pull request as ready for review April 16, 2026 09:27
@piojanu piojanu requested review from a team as code owners April 16, 2026 09:27
@piojanu piojanu changed the title feat: enforce mandatory post-run monitoring in launching-evals skill feat: improve launching-evals and nel-assistant skills Apr 16, 2026
piojanu and others added 2 commits April 16, 2026 11:31
Make step 3 (monitor progress) mandatory after every `nel run` submission.
Agent must use `/loop` to poll until the job reaches a terminal state.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Signed-off-by: Piotr Januszewski <pjanuszewski@nvidia.com>
…skills

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Signed-off-by: Piotr Januszewski <pjanuszewski@nvidia.com>
`/loop` is a Claude Code-specific slash command not understood by
Cursor or OpenAI Codex. Rephrase to "poll status repeatedly" so
the skill works across AI coding tools.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Signed-off-by: Piotr Januszewski <pjanuszewski@nvidia.com>
@chtruong814 chtruong814 added the docs-only With great power comes great responsibility. label Apr 16, 2026
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

docs-only With great power comes great responsibility.

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants