Grow
npx machina-cli add skill Speedy0000007/claude-tandem/grow --openclawTandem Grow
Review your technical profile built automatically from your sessions and identify high-impact skill gaps.
Profile
A single USER.md file at ~/.tandem/profile/USER.md (override directory with TANDEM_PROFILE_DIR).
Three sections:
- Career Context — role, stack, strengths, goals
- Technical Understanding — what you know well (Grow won't repeat these)
- Growth Edges — where you're building depth (nudge targets)
Updated automatically at session end when the session reveals something about understanding level. Kept under 80 lines.
Commands
/tandem:grow (no args)
Show the user's profile:
- Read and display
USER.md - Show line count and last modified date
- If profile is empty or only contains the template, guide the user to fill in Career Context
/tandem:grow gaps
Cross-references five data sources to produce actionable learning priorities:
- Profile (what you know) — from
USER.md - Insights facets (where you have friction) — from
~/.claude/usage-data/facets/*.json. Read files and adapt to whatever fields are present. - Career context (where you want to go) — from the Career Context section of
USER.md - Recurrence (what keeps coming up) — from
~/.tandem/state/recurrence.json - Global activity (what you've been working on across projects) — from
~/.tandem/memory/global.md
Analysis steps:
- Read
USER.mdand categorise by technical domain - Read recent facet files (last 30 days) and aggregate friction by domain
- Read recurrence.json for theme counts
- Read global.md for recent cross-project activity to weight domains by current relevance
- Identify areas where: high friction + thin profile coverage + alignment with goals + high recurrence = strongest signal
- Output a prioritised list: what would make the biggest difference to learn next, and why
Format each gap as:
- Area: [technical domain]
- Signal: [what the friction data shows]
- Recurrence: [theme count from recurrence.json, if applicable]
- Current knowledge: [how much profile coverage exists]
- Recommendation: [specific concept or pattern to learn, with a concrete starting point]
- Impact: [why this matters for their stated goals]
No Profile Yet?
If USER.md doesn't exist or only contains the template comments, guide the user:
- Fill in Career Context with their background and goals
- Explain that the profile builds automatically as they work
- Suggest running
/tandem:grow gapsafter a few sessions of normal usage
Source
git clone https://github.com/Speedy0000007/claude-tandem/blob/main/plugins/tandem/skills/grow/SKILL.mdView on GitHub Overview
Grow helps you review your automatically built technical profile and identify high-impact skill gaps. It combines session data and usage insights to prioritize what to learn next, guiding targeted technical development.
How This Skill Works
Grow reads your USER.md profile (Career Context, Technical Understanding, Growth Edges) and cross-references data sources such as usage facets, recurrence, and global activity. It then outputs a prioritized list of gaps with clear recommendations and expected impact.
When to Use It
- Before a technical interview or code discussion to sharpen focus
- After completing a project to identify new growth areas
- When you want to align learning with your stated goals
- When you invoke /tandem:grow in a session to view your profile
- When you want concrete, data-driven learning priorities (gaps)
Quick Start
- Step 1: Run /tandem:grow to view your profile
- Step 2: Run /tandem:grow gaps to generate prioritized gaps
- Step 3: Pick a top gap and start with the concrete starting point from the Recommendation
Best Practices
- Keep USER.md up to date with Career Context and Growth Edges
- Review facets and recurrence data within 30 days for fresh gaps
- Balance depth in one domain with breadth across your stack
- Turn each gap into a concrete starting point and a micro-goal
- Use gaps to drive your learning plan for the sprint
Example Use Cases
- Run /tandem:grow gaps and see Area: Data Modeling with high friction; start a hands-on data modeling mini-project
- Career Context targets cloud engineering; identify Kubernetes gaps and begin a 2-week repo-based plan
- Recurring API design signals lead you to a pattern-based study guide and practical implementation
- Update USER.md and re-run /tandem:grow gaps to refresh priorities before a job interview
- Global activity highlights cross-project trends; select one domain to deepen with mentoring support
