Improve

skills/j-improve/ is the canonical, hand-edited directory for this skill, per CLAUDE.md's "The Canonical Naming Contract" (the E50 reopening of 2026-09-09, which promoted skills/j-improve/ from generated twin to sole canonical form). The j- prefix is there for collision safety — a real directory under a distinct name, so a host tool shipping its own same-named built-in command cannot shadow it (Claude Code's native skill resolution is a literal-string, directory-name-based match; see project/documentation/skill-authoring.md's "Invocation Convention").

⚠️ scripts/generate-j-alias.sh was retired by E50_S14 and no longer exists — there is nothing to run. This file was previously generated from a bare skills/improve/SKILL.md source; E50_S15 deleted that directory. This file is now the sole canonical, hand-edited source for this skill — edit it directly.

Instructions

Before doing anything, check the attached description for an explicit or implicit goal.

• Explicit goal: clearly stated (e.g. "improve login performance")
• Implicit goal: inferable from context (e.g. "the checkout flow feels slow" → goal: improve checkout performance)

If no goal can be determined, ask:

"What's the outcome you're trying to achieve with this improvement?"

Do not proceed until a goal is confirmed.

Steps

1. Search the Codebase

Search for files and flows relevant to the target goal. Look for:

• Entry points, handlers, or controllers related to the feature
• Shared utilities or services it depends on
• Recent changes in the area (if git history is accessible)

2. Check Project Documentation

Look in project/documentation/ for .md files that offer insight into:

• Architecture decisions
• Known limitations
• Intended behavior of the relevant area

3. Evaluate Insight Sufficiency

Assess whether you have enough context to evaluate current behavior against the goal.

• If yes: proceed to step 5.
• If no: proceed to step 4.

4. Run /examplify

Invoke the /examplify skill. Specify clearly what you want it to find out — e.g. which code paths are exercised, what inputs/outputs look like, or what edge cases exist.

5. Run /evaluate

Copy .agents/skills/j-evaluate/assets/evaluation_invokation_template.yml and fill in:

• paths: path(s) to the relevant example files
• goal: the confirmed target goal

This generates <target-goal-in-kebab-case>-eval.md inside project/rapports/analysis/.

6. Invoke /todo

Call /todo with the following message:

"invoke /brainstorm <path-to-evaluation-rapport>".