promote-skill
Skill25 starsChanged 6 months agoMITView on GitHub
Move a tested personal agent skill into a shared skills repository, check for overlap, normalize metadata, update the catalog and release files, validate the package, and publish it through a reviewable branch. Use when the user asks to promote, publish, share, or contribute a global skill.
In this repository13 other files it gives its agents.
- AGENTS.mdAGENTS.md
- CLAUDE.mdCLAUDE.md
- content-wireframeSkill
- new-projectSkill
- orientSkill
- review-prSkill
- shipSkill
- squash-branchSkill
- release-notesSkill
- writing-ticketsSkill
- hostSkill
- advisorSkill
- sounding-boardSkill
The file
Show its rules and text
We split this skill into the rules it gives an agent
17 rules. About 0 are common, 13 are distinctive, and 4 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈13
Distinctive · estimate
This author's own choices. Worth a look.
4
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 29
Read the destination repository's AGENTS.md, CONTRIBUTING.md, and existing skill examples before making changes.
- Distinctive: line 34
Search personal and repository skills by:
- Distinctive: line 53
Start from an updated default branch and create a focused feature branch:
- Distinctive: line 62
Do not work directly on the default branch unless the user explicitly asks.
- Distinctive: line 73
Use repository metadata conventions.
- Very specific: line 86
Keep SKILL.md focused and under 500 lines.
- Distinctive: line 91
Update every catalog used by the destination repository:
- Very specific: line 93
Add the skill path to .claude-plugin/plugin.json
- Distinctive: line 94
Add the skill to the correct table in README.md
- Very specific: line 95
Add a new entry to CHANGELOG.json
- Distinctive: line 97
Use the plugin manifest version as the release source of truth:
- Distinctive: line 108
Check all of the following:
- Distinctive: line 114
Every listed skill path exists
- Distinctive: line 117
No secrets or personal project data are included
- Distinctive: line 120
Run repository checks when available.
- Distinctive: line 124
Commit with a concise message, push the feature branch, and open a pull request that lists:
- Distinctive: line 132
Do not delete the personal source skill until the repository change is merged and the installed repository version is confirmed.
≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Promote Skill
Promote a tested personal skill into a shared repository without creating duplicate skills or losing its working version.
Required inputs
Identify:
- The source skill directory
- The destination repository
- The destination category
- The repository's default branch and contribution rules
Personal skills commonly live in ~/.cursor/skills/ or ~/.agents/skills/.
Read the destination repository's AGENTS.md, CONTRIBUTING.md, and existing
skill examples before making changes.
1. Check for overlap
Search personal and repository skills by:
- Skill name
- Description and trigger phrases
- Main workflow steps
- Commands and safety rules
Classify similar skills as:
- Duplicate: same purpose and workflow
- Extension: one skill is a more complete version of another
- Adjacent: related but has a different outcome
Recommend merging or replacing duplicates. Keep adjacent skills separate with clear trigger boundaries. Get the user's approval before changing the proposed publication set.
2. Prepare the repository
Start from an updated default branch and create a focused feature branch:
git fetch origin
git checkout <default-branch>
git pull --ff-only origin <default-branch>
git checkout -b feature/add-<skill-name>-skillDo not work directly on the default branch unless the user explicitly asks.
3. Normalize the skill
Copy the complete skill directory, including needed scripts, references, and assets, into:
skills/<category>/<skill-name>/Use repository metadata conventions. For Agent Resources:
---
name: skill-name
description: What the skill does and when it should trigger.
license: MIT
metadata:
author: jcottam
version: "1.0.0"
---Keep SKILL.md focused and under 500 lines. Remove machine-specific paths,
secrets, temporary files, and private project details.
4. Update release files
Update every catalog used by the destination repository:
- Add the skill path to
.claude-plugin/plugin.json - Add the skill to the correct table in
README.md - Add a new entry to
CHANGELOG.json
Use the plugin manifest version as the release source of truth:
- Patch: fixes only
- Minor: new skills or improvements
- Major: breaking changes
The newest changelog version must match the plugin manifest version. Keep skill versions independent in each skill's metadata.
5. Validate
Check all of the following:
- YAML frontmatter parses and contains required metadata
- The directory name matches the skill name
plugin.jsonis valid JSON- The plugin skills array has no duplicate paths
- Every listed skill path exists
- The README link points to the new
SKILL.md - Referenced scripts and files exist
- No secrets or personal project data are included
- The working tree contains only intended changes
Run repository checks when available. Review the final diff before publishing.
6. Publish
Commit with a concise message, push the feature branch, and open a pull request that lists:
- Skills added or replaced
- Overlap decisions
- Version changes
- Validation performed
Do not delete the personal source skill until the repository change is merged and the installed repository version is confirmed.
Final report
Report:
- Published skill path
- Replaced or retired skills
- Manifest and changelog version
- Validation results
- Pull request URL
- Any local cleanup still needed
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

