Generates prioritized interactive improvement plans for AI literacy levels L0-L5, mapping repo gaps to Claude Code plugins/skills with accept/skip/defer after assessment or known level.
How this skill is triggered — by the user, by Claude, or both
Slash command
/ai-literacy-superpowers:literacy-improvementsThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
Generate a prioritised improvement plan that maps assessment gaps to
Generate a prioritised improvement plan that maps assessment gaps to specific plugin commands and skills. Each improvement is presented interactively — the user chooses to accept, skip, or defer.
This skill is invoked by /assess after Phase 5 (workflow
recommendations), but can also be used standalone when the user
knows their current level.
The skill needs two things:
If invoked from /assess, the level is already known — skip this step.
If invoked standalone, ask the user:
What is your current AI literacy level?
If you don't know, run /assess first — it takes fifteen minutes
and gives you an evidence-based answer.
1. Level 0 — Awareness
2. Level 1 — Prompting
3. Level 2 — Verification
4. Level 3 — Habitat Engineering
5. Level 4 — Specification Architecture
Level 5 teams do not need this skill — they are already at the top.
Present the levels above the current one:
You're currently at Level N (Level Name).
How far would you like to improve?
1. Level N+1 — [name] (recommended next step)
2. Level N+2 — [name]
...up to Level 5
The default recommendation is always the next level. Higher targets include all intermediate levels — choosing L4 from L2 means doing L2→L3 improvements first, then L3→L4.
Read the mapping from references/improvement-mapping.md. For each
level transition between current and target:
Check existing state — for each gap in the mapping, verify whether it is already closed. Check the file system:
ai-literacy-assessment
skill as guidance.Filter to open gaps — remove items where the file or configuration already exists and is active.
Assign priority using this heuristic:
Group by level transition — "To reach Level 3" then "To reach Level 4".
Order within each group — high priority first, then medium, then low.
For each improvement, present one at a time:
Improvement 1/N (Level M — Level Name):
Gap: [what is missing]
Action: Run [command] or use [skill]
Priority: High — [one-sentence rationale]
Accept / Skip / Defer?
Handle each response:
If executing a command produces further interactive prompts (e.g.,
/harness-init asks about features), let them run naturally. Resume
the improvement plan after the command completes.
If an assessment document exists for today
(assessments/YYYY-MM-DD-assessment.md), append an Improvement Plan
section:
## Improvement Plan
- Current level: LN
- Target level: LM
- Improvements accepted: N
- Improvements skipped: N
- Improvements deferred: N
- Commands executed: [list of commands/skills that ran]
### Accepted
| Gap | Action | Result |
| --- | --- | --- |
| No HARNESS.md | /harness-init | HARNESS.md created with 4 constraints |
### Skipped
| Gap | Reason |
| --- | --- |
| No Docker scanning | No Docker in this project |
### Deferred
| Gap | Action | Reason |
| --- | --- | --- |
| No fitness functions | fitness-functions skill | Team wants to stabilise L3 first |
If no assessment document exists for today, write the plan to
assessments/YYYY-MM-DD-improvements.md as a standalone document.
Print a summary:
Improvement plan complete.
Current level: L2 (Verification)
Target level: L3 (Habitat Engineering)
Accepted: 4 improvements
Skipped: 1
Deferred: 1
Commands executed: /harness-init, /harness-constrain, /reflect, /harness-health
Run /assess again in 3 months to measure progress.
When used outside /assess, the skill follows the same process but
starts from Step 1 (confirm level). The user can say:
| Priority | Criteria | Examples |
|---|---|---|
| High | Foundational — other items at this level depend on it | HARNESS.md for L3, CI pipeline for L2, specs dir for L4 |
| Medium | Valuable — closes a real gap independently | Secret scanning, GC rules, convention extraction |
| Low | Conditional — depends on project context or gap is partial | Docker scanning (no Docker), convention sync (one AI tool) |
Items marked as "(manual — outside plugin scope)" in the mapping are presented as guidance rather than executable actions.
npx claudepluginhub habitat-thinking/ai-literacy-superpowers --plugin ai-literacy-superpowersAssesses team's AI collaboration literacy by scanning repo for signals like habitat docs, CI workflows, vulnerability scans, tool configs; generates report and badge.
Reads improvement plans generated by skill-quality-reviewer and automatically applies changes to Claude Skills. Useful for executing structured improvements from quality reviews.
Routes leaders discussing AI strategy, tools, or team adoption to relevant skills via fluency assessment and context matching.