Auto Updater
Run a gated five-step workflow to batch-apply safe improvements to multiple SKILL.md packages after reviews, with backups and rollback.
Overview
auto-updater is a journey-wide agent skill that safely applies systematic improvements across multiple skills from review and best-practice inputs—usable whenever a solo builder needs to refresh SKILL.md quality before t
Install
npx skills add https://github.com/adaptationio/skrillz --skill auto-updaterWhat is this skill?
- Five-step workflow: identify improvements, assess safety, apply updates, validate, rollback on failure
- Separates auto-safe changes from changes that need human approval
- Applies updates with backups before touching skill trees
- Post-change validation to ensure quality did not regress
- Explicit rollback path when automated application fails
- 5-step auto-update workflow: identify, assess safety, apply, validate, rollback
Adoption & trust: 1 installs on skills.sh; 11 GitHub stars; 2/3 security scanners passed (skills.sh audits); trending (+100% hot-view momentum).
What problem does it solve?
You have review findings and learned patterns spread across dozens of skills but no safe, repeatable way to apply fixes at scale without breaking working packages.
Who is it for?
Skill authors and indie catalog maintainers running periodic hygiene passes after system-reviewer or best-practices-learner output.
Skip if: First-time authoring of a brand-new skill (use brainstorming and writing-plans first) or one-off manual edits where automation overhead exceeds benefit.
When should I use this skill?
Applying systematic improvements, automating enhancement cycles, bulk updating multiple skills, or implementing ecosystem-wide improvements from review and best-practice findings.
What do I get? / Deliverables
Validated, backed-up skill updates land across the ecosystem, with failures rolled back instead of leaving half-edited SKILL.md trees.
- Patched skill files with pre-change backups
- Validation report of applied vs skipped changes
- Rollback restore point when validation fails
Recommended Skills
Journey fit
Useful at every journey phase - explore requirements and options before committing to a direction.
Where it fits
Normalize frontmatter and allowed-tools blocks across a new skill monorepo before first publish.
Apply auto-safe wording fixes from a code-and-skill review pass without touching risky behavioral sections.
Run a quarterly improvement cycle with backups after best-practices-learner emits new cross-skill patterns.
Refresh FAQ and trigger lists on many marketing-oriented skills after SEO review.
Batch-update install docs and trigger phrases before announcing a skills pack release.
How it compares
Meta maintenance workflow for skill repos, not a single-domain integration like a quote or video API skill.
Common Questions / FAQ
Who is auto-updater for?
Solo and indie builders who maintain multiple agent skills and want automated, validated bulk updates instead of hand-editing every SKILL.md after each review.
When should I use auto-updater?
After Ship or during Operate when review cycles surface repeatable fixes; during Build when normalizing a new skill pack before publish; during Grow when content skills need SEO or structure refreshes—always when improvements are already identified and classified.
Is auto-updater safe to install?
The skill is designed around safety classification, backups, validation, and rollback, but it can write and run shell commands—review the Security Audits panel on this Prism page and restrict allowed-tools in untrusted environments.
SKILL.md
READMESKILL.md - Auto Updater
# auto-updater **Automated improvements** - safely apply enhancements across multiple skills. ## Quick Start 5-Step Workflow: 1. Identify Improvements - Gather recommendations 2. Assess Safety - Classify auto-safe vs manual 3. Apply Updates - Implement with backups 4. Validate Changes - Ensure quality maintained 5. Rollback if Needed - Revert failures **Result**: Safe, validated, automated improvement application --- **Enables safe automated enhancement across ecosystem.** --- name: auto-updater description: Automatically apply improvements to skills and the ecosystem based on system-reviewer findings and best-practices-learner insights. Workflow for automated improvement identification, priority assessment, safe application, validation, and rollback capability. Use when applying systematic improvements, automating enhancement cycles, bulk updating multiple skills, or implementing ecosystem-wide improvements. allowed-tools: Read, Write, Edit, Glob, Grep, Bash, WebSearch, WebFetch --- # Auto Updater ## Overview auto-updater automatically applies improvements to skills and ecosystem components based on identified patterns and learnings. **Purpose**: Automated application of validated improvements across ecosystem **The 5-Step Auto-Update Workflow**: 1. **Identify Improvements** - Gather recommendations from reviews and learnings 2. **Assess Safety** - Determine which can be safely automated 3. **Apply Updates** - Implement improvements automatically 4. **Validate Changes** - Ensure improvements effective, no regressions 5. **Rollback if Needed** - Revert changes if validation fails **Safety**: Always validates before finalizing, can rollback ## When to Use - Applying systematic improvements across multiple skills - Implementing guideline updates ecosystem-wide - Automating common enhancement patterns - Bulk updates (e.g., add Quick Reference to all skills missing it) ## Auto-Update Workflow ### Step 1: Identify Improvements **Sources**: - system-reviewer recommendations - best-practices-learner documented patterns - review-multi common findings - Manual improvement requests **Output**: List of potential improvements **Time**: 15-30 minutes --- ### Step 2: Assess Safety **Safe to Automate**: - Structural additions (add Quick Reference section) - Content additions (add examples in standard locations) - Format standardization (consistent heading levels) - Documentation updates (README enhancements) **NOT Safe to Automate**: - Logic changes (requires understanding context) - Content rewrites (needs judgment) - Major refactoring (risk too high) - Custom implementations **Output**: Classified improvements (auto-safe vs manual-only) **Time**: 20-40 minutes --- ### Step 3: Apply Updates **Process**: 1. Backup affected skills (git commit or copy) 2. Apply improvement to each skill 3. Log changes made 4. Track success/failure per skill **Approach**: One skill at a time, validate each before moving to next **Time**: Varies by improvement and skill count --- ### Step 4: Validate Changes **For Each Updated Skill**: 1. Run skill-validator (pass/fail) 2. Run review-multi structure check (score maintained?) 3. Visual inspection (looks correct?) 4. Mark as validated or flagged for review **Output**: Validation results per skill **Time**: 10-15 minutes per skill --- ### Step 5: Rollback if Needed **If Validation Fails**: 1. Identify which skill failed 2. Restore from backup (git revert or copy back) 3. Analyze why it failed 4. Mark improvement as manual-only for that skill **Output**: Rolled back skill, failure analysis --- ## Example Auto-Update ``` Auto-Update: Add Quick Reference to All Skills Missing It Step 1: Identify - Improvements: Add Quick Reference section - Target Skills: planning-architect, task-development, todo-management - Count: 3 skills to update Step 2: Assess Safety - ✅ Safe: Adding new section (doesn't modify existing content) - ✅ Safe: Standard format (use template) - ✅ Safe: Low