plugin-submission
Skill13 starsChanged 27 days agoApache-2.0View on GitHub
Govern discovery, eligibility, and submission of a plugin or skill to official directories, registries, marketplaces, and curated GitHub lists. Use when preparing or sending plugin listings, marketplace forms, directory PRs, or repository recommendations; require current policy evidence and a final human confirmation before every external submission.
In this repository26 other files it gives its agents.
- AGENTS.mdAGENTS.md
- CLAUDE.mdCLAUDE.md
- adk-expertSkill
- adversarial-gateSkill
- cloud-checkupSkill
- cloud-expertSkill
- cost-guardrailSkill
- cxo-oversightSkill
- delivery-orchestratorSkill
- domain-validatorSkill
- frontend-performance-auditSkill
- get-it-doneSkill
- github-managerSkill
- governance-guardrailSkill
- mcp-server-scaffoldSkill
- operating-model-bootstrapSkill
- product-ownerSkill
- pr-reviewerSkill
- release-managerSkill
- release-readinessSkill
- sitrepSkill
- spec-first-deliverySkill
- system-performance-guardrailSkill
- test-strategySkill
- the-architectSkill
- using-the-harnessSkill
The file
We split this skill into the rules it gives an agent
6 rules. About 0 are common, 5 are distinctive, and 1 is very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈5
Distinctive · estimate
This author's own choices. Worth a look.
1
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 8
Treat every public listing as an external representation of the project.
- Very specific: line 17
Read the current primary policy: Use the destination's official submission instructions or CONTRIBUTING.md.
- Distinctive: line 20
Validate eligibility: Check the public candidate URL, release/tag, licence as detected by the hosting platform, manifest/schema validity, …
- Distinctive: line 33
Send through the permitted route: Use a form, PR, issue, or API only when the policy permits that route.
- Distinctive: line 38
If a listing is rejected or stale, record the reason and next eligible action.
- Distinctive: line 44
Do not send until all applicable checks pass:
≈ “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
Plugin Submission
Treat every public listing as an external representation of the project. Submit only an artifact that the destination's current policy accepts, and retain a receipt for every send.
Workflow
- Classify the destination and artifact. Determine whether it accepts a full
plugin, a repository, or one standalone
SKILL.md. Do not describe a multi-skill plugin as one skill, or split a skill bundle without an explicit portable artifact. - Read the current primary policy. Use the destination's official submission
instructions or
CONTRIBUTING.md; record its URL, access requirements, licence and validation rules, and any limits on automated or AI-assisted submissions. - Validate eligibility. Check the public candidate URL, release/tag, licence as
detected by the hosting platform, manifest/schema validity, documentation, and the
destination-specific validator. Treat a mismatch or unmet requirement as
blocked, not as an invitation to bypass it. - Prepare one destination matrix. For each target, record: destination, artifact,
policy evidence, status (
ready,blocked,deferred, orhuman-only), exact payload, and rollback action. Reuse the canonical repository description; minimise personal data and never add an email without explicit permission. - Stage, then confirm. Fill or draft public data only after the operator has approved the destination. Immediately before clicking a submit control, opening a PR, or sending a recommendation, restate the exact target and payload and obtain a final human confirmation. A single confirmation may cover multiple named sends only when the full set and payloads are shown together.
- Send through the permitted route. Use a form, PR, issue, or API only when the policy permits that route. Never act where a policy requires a human-authored recommendation, prohibits AI-assisted submissions, or requires credentials/access the operator has not supplied.
- Verify and record the receipt. Capture the submission URL/identifier and status. If a listing is rejected or stale, record the reason and next eligible action. Roll back a repository contribution by closing its PR/issue when appropriate; external directory records require the destination's own removal/update path.
Submission Gate
Do not send until all applicable checks pass:
- The target policy is current and supports the proposed artifact.
- The public repository and exact release are accessible.
- Project licence intent and platform-detected licence agree.
- Required validation has passed, including the destination's validator where available.
- The project has not overstated supported platforms, usage, security review, or affiliation.
- The operator has given final confirmation for the exact outgoing entries.
Outputs
- A destination matrix with evidence URLs and eligibility verdicts.
- A staged payload for every ready target.
- Submission receipts, or an explicit blocked/deferred reason and owner.
Anti-rationalisation
| Temptation | Gate |
| “It is close enough to call the bundle one skill.” | Submit the artifact type the directory actually accepts, or create a deliberate portable export first. |
| “The README says Apache, so platform metadata does not matter.” | Directories evaluate the public repository; resolve detected-licence mismatches before broad listing. |
| “The user asked to submit, so the final click is implied.” | A final confirmation binds authority to the exact external payload. |
| “A PR is harmless discovery.” | A public PR is a durable representation of the project and must follow the contributor policy. |
| “We can automate a human-only form.” | Human-only and no-AI policies are hard stops. |
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.