agentsclimarketplace

Sf client discovery

Skill SahirVhora/sf-agent-skills/skills/sf-client-discovery

AI skills for SAP SuccessFactors consultants: configuration health, migration readiness, and HR transformation workflows

Install
npx -y skills add SahirVhora/sf-agent-skills --skill sf-client-discovery

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

One thing to look at

  • 0 stars0 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

What its author says it does

Copied from the file, not written here

Use when you need to generate structured discovery questions for a new or existing SF client -- adapted to what is already known about their tenant.

The file declares its own license as MIT. That is the author’s claim about this one file, and it is not the same thing as the license GitHub reports for the repository, which is listed with the other numbers below.

SKILL.md

3.7 KB, as published. Nobody here has run it

SF Client Discovery

Generate structured, audience-appropriate discovery questions that uncover the real configuration pain points.

When to Use

  • First client engagement (greenfields discovery)
  • Pre-workshop preparation
  • Expanding scope with an existing client
  • Building a proposal or SOW
  • Client has configuration issues but can't articulate them

Prerequisites

  • Optional: previous scan results, tenant metadata summary, client briefing notes
  • Knowledge of client's SF modules in use (EC, Compensation, Recruiting, etc.)
  • Client context (size, countries, merger history, known pain points)

Workflow

Step 1: Load Context

If scan results exist, extract key themes to tailor questions. If no data, start with broad discovery categories.

Step 2: Generate Questions

Use the MCP assessment tool for config-debt-specific questions:

mcp__sf_config_debt_scanner__sf_assessment_questions(category="all")

Supplement with module-specific questions:

  • Governance: Who owns config changes? Change approval process? Audit trail?
  • Custom Fields: How many? Who creates them? Any field rationalisation done?
  • Business Rules: Who writes and tests rules? Documentation? Rule conflict testing?
  • Foundation Objects: How often are hierarchies updated? Who owns them?
  • MDF: Custom objects in use? Who maintains them? Integration dependencies?
  • Integrations: How many? Middleware? Error handling? Monitoring?

Step 3: Pain-Point Prioritisation

Ask the client to rank pain points. Use forced-choice questions:

  • "If you could fix one thing about your SF setup tomorrow, what would it be?"
  • "What takes your HRIS team the most time each week?"
  • "What do users complain about most?"
  • "What broke last time you made a config change?"

Step 4: Package Workshop Agenda

Produce a timed agenda:

  • 15 min: context and goals
  • 30 min: pain-point deep-dive
  • 30 min: guided discovery (question bank)
  • 15 min: executive debrief and next steps

Edge Cases

  • Greenfields (never reviewed): stick to generic assessment questions; don't assume problems
  • Post-migration review: focus on migration-specific pain points (data quality, missing fields, broken integrations)
  • Multi-country rollout: adapt questions per country where config may diverge
  • Post-merger consolidation: ask about both tenants separately, then consolidation plan
  • Non-English client: generate questions in client's language
  • Procurement/legal audience: strip technical detail; focus on risk, compliance, and cost

Common Pitfalls

  1. Leading questions: "Your custom fields are out of control, right?" → "How do you manage custom field creation?"
  2. Too many questions: 30+ questions overwhelm. Top 15, grouped by theme, with optional deep-dive sections.
  3. No success definition: Always ask "what does a successful review look like?" early.
  4. Technical jargon with non-technical audience: HR Directors don't know what MDF or OData means.
  5. Not sending pre-read: Clients who see questions in advance give better answers. Include a 5-question pre-read.

Verification Checklist

  • Questions categorised by governance area
  • Pain-point prioritisation exercise included
  • Success definition captured first
  • Audience-appropriate language used
  • Timed workshop agenda produced
  • Pre-read questions prepared
  • Next-steps decision tree included

Keep looking

Skills are one crate of 328,083. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.