jfrog-dashboard-blueprint - SKILL.md Agent Skill

name: jfrog-dashboard-blueprint description: >- Scaffold a new persona-specific JFrog dashboard/report skill from a guided interview. Use when the user asks to "create a new dashboard", "scaffold a report skill", "build a CTO/Eng/Compliance/DevOps dashboard pack", or "generate a starter blueprint for a custom report". This skill DOES NOT query JFrog APIs or generate a live report. It produces a new skill folder (SKILL.md, schema, data-collection mapping, HTML template, sample fixtures) that a persona-specific reporting skill can be built from. metadata: role: workflow author: Avinash Giri version: 2.1.1

JFrog Dashboard Blueprint Generator

CRITICAL — READ THIS FIRST

You are NOT generating a live report. You are NOT calling JFrog APIs. You are NOT producing HTML for a specific run. Your only job is to scaffold a new persona dashboard skill by:

Running a short interview with the user.
Producing a new skill folder with the standard file layout.
Seeding each file from the templates in references/.
Replacing placeholders with the persona's answers.
Telling the user what to fill in next and how to register the skill.

If the user asks for an actual report run, hand off to the matching persona skill (for example jfrog-ciso-report). Do NOT attempt to run a report from this skill.

When to use this skill

Use when the user wants a starter blueprint for a new dashboard pack:

"Create a new dashboard for our CTO"
"Scaffold an Engineering Head report"
"Build a compliance pack"
"I want a starter blueprint for a custom DevOps dashboard"
"Generate a new persona skill"

Do NOT use for live report generation. Do NOT use for editing an existing persona pack.

Workflow

1. Run the persona interview (references/persona-interview.md)
2. Resolve output path (default: ./dashboard-report-skills-<persona-slug>)
3. Create the folder structure
4. Generate files from templates with placeholders substituted
5. Print a "next steps" summary for the user

Step 1: Persona interview

Open references/persona-interview.md and ask the user the 7 core questions in a single message. Wait for answers before continuing. Do not guess answers. Do not skip questions. If the user gives partial answers, ask only the missing items in a follow-up.

The answers MUST yield these variables:

Variable	Source
`PERSONA_NAME`	Question 1
`PERSONA_SLUG`	Lowercase, hyphenated form of `PERSONA_NAME`
`AUDIENCE`	Question 1 (exec / manager / IC)
`KEY_QUESTIONS[]`	Question 2 (up to 5)
`DECISIONS[]`	Question 3
`DATA_SOURCES[]`	Question 4
`CADENCE`	Question 5 (daily / weekly / monthly)
`OUTPUT_FORMATS[]`	Question 6 (html / email / slack / pdf)
`TRUST_REQS[]`	Question 7 (sources, approvals, audit)

Step 2: Resolve output path

Default: ./dashboard-report-skills-<PERSONA_SLUG>/. Ask only if the default conflicts with an existing folder. If conflict, offer:

Use a suffix (-v2)
Pick a different path
Abort

Step 3: Create folder structure

dashboard-report-skills-<PERSONA_SLUG>/
├── SKILL.md
└── references/
    ├── report-schema.md
    ├── report-data-collection.md
    ├── dashboard.html
    └── sample-data.json

Use the templates in this skill's references/ directory as the source for each file. Substitute placeholders (see Step 4).

Step 4: Placeholder substitution

Every template uses double-curly placeholders. Replace all of them. If a value is missing, replace with TODO: <description> so the user knows what to fill in.

Placeholder	Replace with
`{{PERSONA_NAME}}`	e.g. "Engineering Head"
`{{PERSONA_SLUG}}`	e.g. "eng-head"
`{{AUDIENCE}}`	e.g. "engineering leadership"
`{{KEY_QUESTIONS}}`	Bullet list of the 5 questions
`{{DECISIONS}}`	Bullet list of decisions this report supports
`{{DATA_SOURCES}}`	Bullet list of JFrog products and any external sources
`{{CADENCE}}`	weekly / monthly / etc.
`{{OUTPUT_FORMATS}}`	html, email, etc.
`{{TRUST_REQS}}`	source attribution, approvals, audit log requirements
`{{GENERATED_DATE}}`	Today's ISO date

Step 5: Next-steps summary

After creating files, print a short message to the user with:

The folder path created.
A checklist of what to fill in:
- Map each KEY_QUESTION to a schema section.
- Implement API queries in report-data-collection.md.
- Customize visuals in dashboard.html.
- Add at least one golden fixture in sample-data.json.

How to register the skill globally:

npx skills add ./dashboard-report-skills-<PERSONA_SLUG> -g -y

Reminder: keep the JSON-first contract. The agent must always produce JSON before injecting into the HTML template.

Hard rules

Do NOT call JFrog APIs from this skill.
Do NOT write live data to the generated files.
Do NOT skip the interview.
Do NOT generate HTML or schema content that diverges from the templates in references/. Use them as the source of truth.
Every generated SKILL.md MUST include the JSON-first execution contract (the template enforces this).
Every generated dashboard.html MUST be self-rendering and consume a single __DATA__ placeholder.