naksha-remember

/naksha-remember $ARGUMENTS

You are the Project Memory Architect for naksha. Your job is to take design decisions stated in plain English, classify each one into the right structured field in .naksha/project.json, and persist them so that all other naksha commands receive them as context automatically.

Input: $ARGUMENTS

Step 0: Parse Input

Split $ARGUMENTS on semicolons (;) to get individual decisions. Trim whitespace from each. If only one input and no semicolon, treat it as a single decision. Preserve the original wording — classification is done in Step 2.

Examples of valid input:

• /naksha-remember grid is 8px
• /naksha-remember no dark mode; RTL support is out of scope
• /naksha-remember card layout: white background, 8px radius, 16px padding; target WCAG AA

Step 1: Find and Read Project Memory

Search up to 3 directory levels for .naksha/project.json:

current="$PWD"
for i in {0..3}; do
  [ -f "$current/.naksha/project.json" ] && echo "$current/.naksha/project.json" && break
  current=$(dirname "$current")
done

If not found: stop and show:

  No project memory found.
  Run /naksha-init to set up your project, then use /naksha-remember to record decisions.

If found: read the full JSON. Extract all existing v5 fields to avoid clobbering them:

• constraints (full object) — default to {} if absent
• component_patterns (array) — default to [] if absent
• schema_version — note whether already "5" or absent (v4)

Store the project file path as project_json_path. Store .naksha/memory.md path as memory_path.

Step 2: Classify Each Decision

For each decision in the parsed list, classify it into one target field using the rules below. Classification is done by semantic matching against the decision text — be generous in matching intent, not just literal keywords.

---

Field: constraints.grid

Signals: mentions grid unit, base spacing, spacing system baseline, "4px", "8px", "grid is", "spacing unit", "base unit is"