Skip to content

Maintenance

These workflows keep your learnship installation and project in good shape.

/settings

Interactive configuration editor for .planning/config.json.

/settings

Walks you through each setting with explanations and current values. Safer than editing config.json directly: validates values before writing.

See Configuration for the full schema reference.

/set-profile [quality|balanced|budget]

One-step model profile switch: the fastest way to adjust cost vs. quality.

/set-profile quality    # all agents use large models
/set-profile balanced   # mix of large and medium (default)
/set-profile budget     # all agents use smallest viable models

Only changes model_profile in config.json. For full preset changes (mode, granularity, toggles) use /settings.

/health

Project health check with numeric score — surfaces issues before they become blockers.

/health           # check and report
/health --repair  # check and auto-fix repairable issues

What it checks: - Required planning files (PROJECT.md, ROADMAP.md, STATE.md, config.json) - Config validity and completeness - State/roadmap consistency (current phase exists in roadmap) - Phase directories match roadmap phases - Plans without summaries - Uncommitted changes to planning artifacts

Numeric score (0–100):

Health computes a deterministic score. Start at 100, then deduct per issue:

Issue Deduction
PROJECT.md missing −25
ROADMAP.md missing −20
config.json missing −10
config.json parse error −15
STATE.md missing −10
State/roadmap mismatch −5
Missing config fields −2/field (max −10)
Phase dir missing −4/phase (max −12)
Plan without summary −1/plan (max −5)
Uncommitted .planning/ changes −5

Status bands: - HEALTHY (90–100) ✓ — project artifacts are clean - DEGRADED (70–89) ⚠ — issues present but project is functional - BROKEN (0–69) ✗ — critical artifacts missing or inconsistent

The repair footer shows [N] issue(s) can be auto-repaired (+N points). Run: health --repair.

Auto-repairable issues: - STATE.md missing → regenerate from ROADMAP.md - config.json missing or parse error → reset to defaults - config.json missing fields → add missing fields with defaults

Run this any time you feel the project state might be inconsistent. It's also suggested automatically by /ls when planning artifacts drift.

/cleanup

Archives completed milestone phase directories to keep .planning/phases/ clean.

/cleanup

Moves completed phase directories to .planning/milestones/[version]/phases/. Run this after /complete-milestone if you want a cleaner working directory.

/update

Updates the learnship platform itself to the latest version.

/update

Downloads the latest workflows from github.com/FavioVazquez/learnship and installs them, replacing the current workflow files.

Local customizations

If you've edited any workflow files, run /reapply-patches after updating to restore your changes.

/reapply-patches

Restores local workflow customizations after running /update.

When to use: After /update if you had local modifications to workflow files. Works by re-applying a stored patch diff.

/sync-upstream-skills

Syncs @agentic-learning and @impeccable skills from their upstream repositories.

/sync-upstream-skills
  • Pulls latest agentic-learning from github.com/FavioVazquez/agentic-learning
  • Pulls latest impeccable from github.com/pbakaus/impeccable
  • Preserves local customizations (impeccable dispatcher SKILL.md)
  • Runs the installer to update all platform copies

When to use: When upstream skills have been updated and you want the latest learning techniques or design commands.

/release

Cuts a new learnship release: bumps version, updates changelog, pushes tag, creates GitHub release.

When to use: When you're developing learnship itself and want to publish a new version. Not needed for normal project work.