test(source-control): add eval cases for babysit-prs, babysit-loop and setup

Ouverte
#7,148 1 commentaire 0 réactions 0 personnes assignées Voir sur GitHub

Les mainteneurs répondent en général sous 1 jour

Personne n'a encore pris cette issue.

Évaluation

Difficulté
4/5
Temps estimé
3-5 jours
Accessibilité débutants
48/100
Type d'issue
Fonctionnalité
Clarté
Plutôt claire
Activité
Active
Stack technique
github, shell
Domaine
devtools, testing-qa

Piste de recherche

Read the three skills’ SKILL.md files, the eval cases under plugins/source-control/evals/, the skill criteria at plugins/playbooks/skills/skill-authoring/reference/skill-criteria.md, and the gh/PR fixture work in #6997. Design at least one case for each skill with /evals:design, validate with /evals:validate, then run each case paid with and without the skill. Done when all cases load, avoid real GitHub, and record their with/without results.

Rédigé par le modèle d'indexation à partir du texte de l'issue.

Description

needs-triage priority: medium

Observed

plugins/source-control/evals/ has cases for commit, pull-request, resolve-conflicts and worktree, and none for three of its skills:

  • babysit-prs
  • babysit-loop
  • setup

(check is a read-only prerequisites check and is out of scope here.) Changes to these three skills ship with no behavioral evidence, and the skill criteria (plugins/playbooks/skills/skill-authoring/reference/skill-criteria.md) require eval evidence of no loss before a cut to a skill body, which these skills cannot produce today.

Fix wanted

Design at least one case per skill with /evals:design, covering the behavior each SKILL.md treats as most important, for example:

  • babysit-prs: the safe default tier never resolves threads or merges.
  • babysit-loop: a do-not-merge label or a body hold stops the merge.
  • setup: writes only what the user confirmed.

Both babysit skills read GitHub state, so they need the gh/PR fixture from #6997; build on that rather than calling real GitHub. Validate statically (/evals:validate), then run each case paid with/without and record the delta.

Acceptance: each skill has at least one case that loads, never touches real GitHub, and shows a with/without result.

Depends on: #6997

Langage dominant
Shell
Étoiles
22
Forks
2
Merge moyen
1 h 49 min
PR mergées (30 j)
916

Préparer son environnement

Par où commencer

  1. Lisez l'issue en entier, puis le guide de contribution du projet.
  2. Signalez en commentaire que vous la prenez — cela évite que deux personnes fassent le même travail.
  3. Forkez le dépôt et travaillez sur une branche.
  4. Ouvrez une pull request qui référence le numéro de l'issue.

Autres issues de melodic-software/claude-code-plugins

Toutes les issues de melodic-software/claude-code-plugins

Issues similaires

Plus d'issues Shell/Bash

Recevez les nouvelles issues par e-mail

Un résumé court des issues GitHub adaptées aux débutants.