aksheyw You are updating the customization memory — the learning log that makes every future application smarter.
STEP 1: Read the current memory file
Read ${CLAUDE_PLUGIN_ROOT}/references/CUSTOMIZATION_MEMORY.md fully (the user's real, git-ignored file). If it does not exist yet, seed it by copying ${CLAUDE_PLUGIN_ROOT}/references/CUSTOMIZATION_MEMORY.example.md to that path first. Understand the existing patterns before adding anything.
STEP 2: Gather information
Ask the user (or extract from their input):
- Which company and role?
- What was the outcome? (Resume screen passed / rejected / screened / interviewed / offered)
- What resume strategy was used? (Which company-type template?)
- What was the cover letter hook?
- Was there a referral involved?
- For rejections: any feedback or signal about why?
- For successes: what specific framing or metric seemed to resonate?
STEP 3: Identify the learning
Based on the outcome, generate a specific, actionable learning:
- NOT "good resume" — say WHAT specifically seemed to work
- NOT "bad fit" — say which framing missed the mark
- Connect to company-type patterns already in the file
STEP 4: Update the file
Write the updated CUSTOMIZATION_MEMORY.md to ${CLAUDE_PLUGIN_ROOT}/references/CUSTOMIZATION_MEMORY.md.
Add the new entry under the correct company-type section. Update:
- "Successful Customizations" if outcome was positive
- "Unsuccessful Patterns" if outcome was negative or no response
- "Screen Rate" with the updated percentage if calculable
STEP 5: Surface patterns
After updating, analyze across all logged applications:
- Which company types are yielding screens?
- Which cover letter hooks got the most traction?
- Which resume strategies need improvement?
- Are there any keywords or framings that appear in multiple successful applications?
Output a brief "What We Know So Far" summary with 3-5 actionable insights.