Learn

Save a learning to project memory

$ARGUMENTS

Determine the learning

If triggered automatically (multi-cycle resolution), synthesise the learning from the conversation:

• What was the root cause of the extra cycle(s)?
• What assumption or gap in knowledge caused the first attempt to fail?
• What is the correct approach / API / behaviour?
• What should be checked first next time to avoid the same detour?

If triggered by the user, record exactly what they stated in $ARGUMENTS.

Steps

1. Read the current memory file: /Users/alex.mondshain/.claude/projects/-Users-alex-mondshain-claude-jhelm/memory/MEMORY.md

2. Check whether a relevant topic file already exists in that same directory (e.g. dependencies.md, testing.md, debugging.md). If so, read it too.

3. Decide where to write:

   • Short, self-contained insight that fits an existing MEMORY.md section → add it there (keep file ≤ 200 lines).
   • Detailed or topic-specific learning → append to or create a dedicated topic file, then add/update a one-line reference in MEMORY.md.

4. Write in concise, actionable form:

   • Bullet points, not prose.
   • Lead with what to do / what to check, follow with why.
   • If it supersedes an existing note, update or remove the old one.

5. Scan for reusable scripts (see section below).

6. Confirm to the user what was saved and where (one line is enough).

What to save

• Root causes of multi-cycle failures and the correct fix
• Non-obvious library behaviours or API quirks discovered during the task
• Version constraints and compatibility issues (e.g. BCrypt prefix $2y$ vs $2a$)
• Architectural decisions and their rationale
• Workflow or tool preferences the user has stated
• Patterns confirmed to work that aren't obvious from the code
• Platform-specific shell/CLI argument failures — save to the ## macOS / Platform-Specific Shell Quirks section in MEMORY.md with the failing invocation and the correct macOS alternative

What NOT to save

• Routine outcomes that worked first time
• Temporary task state or in-progress work
• Anything already covered verbatim in CLAUDE.md
• Guesses or conclusions that were not verified by a passing test or explicit confirmation

Scan for reusable scripts (token-saving step)

After saving the main learning, check whether any inline scripts generated during the current session should be promoted to persistent memory to avoid regenerating them in future sessions.

When to do this scan: Any time the session generated a Python or shell script of 10+ lines inline (in the conversation, not already in a skill file).

How:

1. Look at the current conversation for python3 -c, python3 /tmp/, or multi-line shell scripts that were written ad-hoc.

2. For each script found, ask:

   • Is this script already embedded in a skill (e.g. /checkstyle, /jacoco)? → Skip.
   • Was this script written once for a one-off task? → Skip.
   • Is this something likely to be needed again (analysis, bulk edits, data parsing)? → Save it.

3. If saving is warranted, choose the destination:

   • Script belongs to an existing skill → embed it in that skill's skill.md under a named section.
   • General utility with no existing skill → save as .claude/scripts/<name>.py (or .sh) in the repository.
   • Complex enough to deserve its own skill → note it in MEMORY.md as a candidate for a new skill.

4. Save the script file to .claude/scripts/<descriptive-name>.<ext> with a comment header:

   # Purpose: <one-line description>
   # Run: python3 .claude/scripts/<name>.py [args]
   <script content>
   

5. Add a one-line reference in MEMORY.md under a ## Scripts section (create if absent):

   - `.claude/scripts/<name>.py` — <purpose>
   

6. Commit the new script directly to main (skill-only rule applies to .claude/ changes):

   git add .claude/scripts/<name>.<ext>
   git commit -m "Add reusable <name> script"
   git push
   

Goal: scripts live in the repo, are version-controlled, and can be run directly with a known path — no regeneration, no memory file bloat.