Skill Publisher Pro for Capafy

Build a focused, compliant, commercially ready Agent—from idea to submission.

Skill Publisher Pro for Capafy is an independent publisher assistant for creators who want to turn expertise, prompts, workflows, documents, or existing code into a premium Skill-based Agent for the Capafy marketplace.

It goes beyond writing a prompt. It designs the product, builds the runtime Skill, prepares the public listing, applies Capafy-aligned guardrails, plans privacy and integrations, recommends monetization, validates the files, and prepares release packages.

Independent tool. Not affiliated with or endorsed by Capafy. Platform requirements may change; the current Publisher Console and official Capafy documents remain authoritative.

Who It Helps

• Skill and Agent creators
• AI consultants
• Prompt engineers
• Developers
• Subject-matter experts
• Agencies packaging internal workflows
• Publishers reviewing or updating existing Capafy Agents

Problems It Solves

Creating a sellable Agent requires more than a SKILL.md. Publishers must align product scope, runtime architecture, listing copy, testing, privacy, policy boundaries, pricing, brand assets, and packaging.

This Agent helps prevent common failures such as:

• generic or weak product positioning;
• bloated runtime packages;
• incomplete Agent Cards;
• unsupported or restricted functionality;
• missing disclaimers;
• inaccurate third-party data declarations;
• untested scripts;
• invented pricing assumptions;
• mismatched public claims and runtime behavior;
• missing runtime or submission files.

Core Capabilities

New Agent Build

Turn a product idea into:

• product and audience definition;
• viability assessment;
• policy classification;
• use cases and workflow;
• production SKILL.md;
• scripts, references, schemas, templates, and workflows when needed;
• Agent Card and README;
• Test Input and Expected Output;
• monetization and pricing;
• privacy and third-party disclosure;
• logo and Hero Image;
• changelog and release checklist;
• runtime and submission ZIPs.

Existing Skill Audit

Review:

• functionality;
• architecture;
• prompt injection resistance;
• security;
• data flows;
• policy fit;
• sensitive-sector restrictions;
• listing accuracy;
• pricing logic;
• release readiness.

Compliance Remediation

Identify and correct:

• prohibited capabilities;
• missing disclaimers;
• unsafe or overconfident outputs;
• unsupported guarantees;
• autonomous irreversible actions;
• HR discrimination risk;
• high-stakes advice boundaries;
• hidden data sharing;
• deceptive synthetic-content use.

Listing & Pricing

Create buyer-facing:

• title;
• description;
• README;
• Welcome Message;
• five tags;
• pricing ladder;
• test case;
• data disclosure;
• positioning and differentiation.

Version Update

Prepare safe updates with:

• change analysis;
• semantic version;
• revised runtime files;
• updated tests;
• updated public claims;
• updated privacy disclosures;
• new release packages.

Inputs

Start with any of these:

• a product idea;
• a workflow or prompt;
• an existing Skill folder or ZIP;
• source code;
• a document describing the desired Agent;
• Publisher Console screenshots;
• current Capafy field requirements.

For stronger results, include:

• target buyer;
• problem solved;
• desired output;
• required files or data;
• expected usage frequency;
• integrations;
• marketplace language;
• preferred monetization;
• whether source-code delivery is acceptable.

Outputs

Depending on scope, the Agent can produce:

• SKILL.md
• agents/openai.yaml
• scripts and tests
• reference documents
• schemas and templates
• workflow files
• Capafy Agent Card
• marketplace README
• description
• Welcome Message
• five tags
• Test Input
• Expected Output
• pricing and revenue logic
• third-party data disclosure
• changelog
• logo
• Hero Image
• validation report
• runtime ZIP
• submission ZIP

How It Works

1. Understand the product and buyer.
2. Score commercial viability.
3. Classify policy and sector risk.
4. Define representative use cases.
5. Choose the smallest reliable architecture.
6. Build and test the runtime.
7. Create the listing and pricing.
8. Map integrations and data flows.
9. Generate and validate brand assets.
10. Validate and package the release.

Example Prompts

Build a new Agent

Create a premium Capafy Agent for marketing teams that analyzes uploaded campaign reports and produces an executive summary, anomaly analysis, prioritized actions, and a client-ready Markdown report.

The marketplace listing should be in English. I prefer Subscription. Do not use external APIs.

Audit an existing Skill

Audit the attached Skill for Capafy review readiness. Check runtime behavior, policy boundaries, data disclosure, security, field limits, pricing, tests, and listing accuracy. Produce a prioritized remediation plan and updated package.

Improve positioning and pricing

Review this Agent concept and decide whether it should use Subscription, hourly, or Download. Create a premium positioning, price ladder, Agent Card, test case, and revenue logic. Mark any Sandbox Fee assumptions as provisional.

Update a version

Update the attached Agent from version 1.1.0. Add multi-file comparison, preserve existing behavior, update tests and disclosures, choose the correct semantic version, and package the release.

Policy-First Design

The Agent classifies proposed products as:

• standard supported;
• sensitive supported with restrictions;
• currently unsupported;
• permanently restricted.

It does not help disguise unsupported or prohibited functionality.

For sensitive sectors, it requires prominent disclaimers, explicit runtime boundaries, refusal behavior, data disclosure, and adversarial tests.

Privacy

This release does not hardcode an additional third-party API, storage service, analytics provider, email service, automation platform, or webhook.

User prompts and files are processed by the Capafy runtime and its disclosed model provider. If the deployed configuration adds another provider, the listing must be changed to disclose that provider, purpose, and data flow before publication.

The Agent minimizes transmitted data and never packages credentials.

Pricing Guidance

The Agent recommends one primary monetization model based on buyer behavior, usage frequency, IP exposure, infrastructure, and retention.

For Subscription, it creates a complete Daily, Weekly, and Monthly configuration: price, request limit, estimated workflows, Weekly/Monthly trial decision, trial duration, free request limit, one featured plan, and net logic. It uses current Publisher Console minimums and Sandbox Fees and does not invent those values.

Limitations

• Capafy requirements can change.
• Publisher Console minimums and Sandbox Fees must be confirmed before publication.
• Generated code must be tested in the target runtime.
• The Agent cannot guarantee Capafy approval.
• Legal, financial, medical, tax, employment, privacy, and regulatory decisions may require qualified review.
• Logo and Hero Image requirements must be confirmed in the current Console.
• The publisher remains responsible for the final submission and ongoing compliance.

Version

1.1.0