skills/skill-sync-checker/SKILL.md
Detects content drift between skill files and their source documents. Helps maintain skills that are derived from other documentation by comparing content and flagging outdated sections.
npx skillsauth add kanyun-inc/reskill skill-sync-checkerInstall this skill globally with one command. Works with Claude Code, Cursor, and Windsurf.
3 of 9 scanners reported clean
Some scanners were skipped, did not run, or reported a non-clean status. Review each row below.
A utility skill that detects when skill content has drifted from its source documents. Any skill file that includes a <!-- source: path/to/file --> marker can be checked for freshness and updated accordingly.
Use this skill when the user:
This skill is automatically injected into context when the user edits README.md (configurable via .cursor/rules/skill-sync-checker.mdc globs).
When auto-triggered, follow this lightweight flow:
skills/**/SKILL.md for <!-- source: ... --> markers referencing the file being editedExample notification:
Note: skills/reskill-usage/SKILL.md is derived from README.md (last synced: 2026-02-12).
If your changes affect CLI commands, options, or usage examples, the skill may need updating.
Run a sync check when you're done editing.
Skill files that are derived from other documents should include source markers as HTML comments at the top of the file (before or after the YAML frontmatter):
<!-- source: README.md -->
<!-- synced: 2026-02-11 -->
---
name: my-skill
description: ...
---
# My Skill
...
| Marker | Required | Description |
| ------------------------- | -------- | ---------------------------------------------------- |
| <!-- source: <path> --> | Yes | Relative path to the source file (from project root) |
| <!-- synced: <date> --> | No | ISO date of last sync (YYYY-MM-DD) |
A file can have multiple source markers if it derives from several documents:
<!-- source: README.md -->
<!-- source: docs/cli-spec.md -->
<!-- synced: 2026-02-11 -->
---), markers can go before or after itWhen asked to check a skill for sync status, follow these steps:
Scan all .md files in the target skill directory. For each file, look for <!-- source: ... --> comments. Files without source markers are original content and can be skipped.
For each source marker, read the referenced file. If the source file does not exist, report it as an error — the source may have been moved or deleted.
Compare the skill file against its source document(s). Focus on structural and factual differences, not formatting:
Key things to check:
Ignore:
Present findings in a clear summary:
Sync Check: skills/reskill-usage/SKILL.md
Source: README.md
Last synced: 2026-02-11
Status: ⚠ Out of sync
Differences found:
1. New command `find` added in source (missing from skill)
2. Option `--registry` added to `install` command
3. New agent "Gemini" added to Multi-Agent Support table
Recommendation: Update the skill to reflect these changes.
If everything is in sync:
Sync Check: skills/reskill-usage/SKILL.md
Source: README.md
Last synced: 2026-02-11
Status: ✓ In sync
No significant differences found.
If the <!-- synced: ... --> marker is absent, report "Last synced: unknown":
Sync Check: skills/example/SKILL.md
Source: docs/example-spec.md
Last synced: unknown
Status: ⚠ Out of sync
...
When differences are detected and the user confirms they want to update:
Present each difference with the relevant content from the source file so the user can review what changed.
After user confirmation, update the skill file to reflect the source changes. Preserve the skill's structure and agent-specific content (When to Use, Workflows, Troubleshooting) — only update the sections that correspond to source material.
Update the <!-- synced: ... --> marker to today's date:
<!-- synced: 2026-02-12 -->
When checking a skill, also scan the project for new documentation that might be relevant but is not yet tracked:
| File Pattern | Likely Relevant | Action |
| --------------------------------------- | ------------------------ | ------------------------ |
| *-spec.md, *-summary.md | Yes — user-facing specs | Suggest adding as source |
| README.md, README.*.md | Yes — user-facing docs | Suggest adding as source |
| API.md, docs/*.md | Yes — user-facing docs | Suggest adding as source |
| CHANGELOG.md | No — auto-generated | Skip |
| CONTRIBUTING.md | No — contributor guide | Skip |
| *-design.md | No — internal design | Skip |
| *-plan.md, *-implementation-plan.md | No — internal plans | Skip |
| *-migration-plan.md | No — internal migration | Skip |
| *-adaptation.md | No — internal adaptation | Skip |
When a potentially relevant new document is found, notify the user:
New document detected: docs/new-feature-spec.md
This looks like a user-facing spec that might be relevant to the skill.
Would you like to review it for inclusion?
Never automatically add new sources or modify skill content without user confirmation.
Here is a complete example of checking the reskill-usage skill:
User: "Check if the reskill-usage skill is up to date"
Agent steps:
1. Read skills/reskill-usage/SKILL.md
2. Find marker: <!-- source: README.md -->
3. Read README.md
4. Compare commands table, source formats, configuration, etc.
5. Report: "The skill is missing the new `find` command that was
added to README.md. Would you like me to update it?"
After user confirms:
Agent steps:
1. Add `find` command to the Commands table in SKILL.md
2. Update <!-- synced: 2026-02-12 -->
3. Report: "Updated SKILL.md — added `find` command, synced date
updated to 2026-02-12"
To check all skills in a project at once:
User: "Check all skills for sync status"
Agent steps:
1. Find all SKILL.md files in the project (skills/, .cursor/skills/, etc.)
2. For each file with source markers, run the detection workflow
3. Present a summary of all results:
Sync Check Summary:
✓ skills/find-skills/SKILL.md — no source markers (original content)
⚠ skills/reskill-usage/SKILL.md — 2 differences found (source: README.md)
✓ skills/skill-sync-checker/SKILL.md — no source markers (original content)
tools
Teaches AI agents how to use reskill — a Git-based package manager for AI agent skills. Covers CLI commands, install formats, configuration, publishing, and common workflows.
development
Assesses code comprehensibility and maintainability risk. Use when the user asks about code confidence, risk, maintainability, tech debt, code health, or whether code is safe to change. Also use when the user asks to analyze code quality, scan for risks, check if code is messy or complex, audit code, do a code checkup, find weak spots, assess what needs refactoring, or asks about code trust, hidden risks, gotchas, or onboarding to a codebase.
development
Behavioral modifier for AI coding assistants working with non-developers. Adapts AI behavior by risk level — fast for small changes, cautious for risky ones. Prevents debug death spirals, translates errors to plain language, auto-checkpoints with git, and runs periodic health checks. Always active, zero manual trigger needed.
data-ai
Helps users discover and install agent skills when they ask questions like "how do I do X", "find a skill for X", "is there a skill that can...", or express interest in extending capabilities. Uses reskill as the package manager.