Salesforce org assessment skill

Salesforce Org Assessment Skill

Purpose

This skill provides a structured workflow for assessing a Salesforce org's overall posture from sanitized metadata exports. It produces a risk register and remediation backlog covering the object model, automation inventory, permission topology, integration map, and technical debt indicators. It does not access live orgs, request credentials, or authorize changes.

When to use

• An org is being evaluated for acquisition, merger, or re-platforming.
• A new technical lead needs a structured health baseline.
• A security or compliance review requires a documented posture assessment.
• A risk register must be produced before a major release or migration.

When not to use

• Single metadata type deep-dive — use salesforce-metadata-review-skill.
• Permission-only review — use salesforce-permission-model-review-skill.
• Live production change proposal — use salesforce-live-change-approval-protocol.
• Automation-only review — use salesforce-flow-automation-review-skill.

Minimum payload (required inputs)

• Sanitized metadata exports (one or more of): object definitions, field lists, automation inventory (Flow list, validation rules), permission set/profile summary, installed packages list, integration endpoints (sanitized, no tokens).
• Org environment type (production / sandbox — use placeholder, not real org ID).
• Business context: industry vertical, approximate user count range, primary use cases (Sales, Service, Marketing, CPQ, etc.).

Workflow

1. Object model review — Identify custom objects, junction objects, deprecated standard objects in use, field count per object (flag objects > configurable threshold), hardcoded IDs in default field values, formula complexity.
2. Automation inventory — List all automation types present (Flow, validation rules, approval processes, triggers, Process Builder ). Flag: mixed automation types on same object, Process Builder still active, missing fault paths, automation running in system context without sharing.
3. Permission topology — Identify profiles, permission sets, permission set groups. Flag: toxic combinations (ModifyAllData, ViewAllData, broad API Enabled without IP restriction). Summarize permission set proliferation.
4. Integration map — List integration patterns present (REST, SOAP, Bulk, Streaming, Platform Events , CDC , middleware). Flag: point-to-point patterns without error queues, missing named credentials, hardcoded endpoints.
5. Technical debt indicators — Assess: deprecated metadata types, unused fields (> configurable threshold with zero usage), hardcoded IDs, legacy APIs still in use, managed package versions behind current release.
6. Risk register — For each finding, assign: risk_tier (from salesforce-risk-taxonomy), matter_type, remediation_priority, owner_role, estimated_effort.
7. Remediation backlog — Ordered list of remediation items with priority, dependency notes, and recommended approach.

Evidence requirements

• Sanitized metadata exports; no credentials, no session tokens, no customer data.
• If exports contain field values that appear to be production data samples, decline and ask for schema-only exports.
• Object and field counts are acceptable; record values are not.

Output format

org_assessment_summary:
  object_model_findings: [list with risk_tier per finding]
  automation_inventory_findings: [list with risk_tier]
  permission_topology_findings: [list with risk_tier]
  integration_map_findings: [list with risk_tier]
  technical_debt_indicators: [list with risk_tier]

risk_register:
  - finding_id: [generated]
    matter_type: [from salesforce-risk-taxonomy]
    risk_tier: Critical | High | Medium | Low | Unknown
    description: [one sentence]
    evidence_basis: [what in the export supports this]
    owner_role: [placeholder role]
    remediation_priority: P1 | P2 | P3 | P4

remediation_backlog:
  - priority: [P1-P4]
    item: [description]
    dependencies: [other items that must precede this]
    recommended_approach: [brief]

escalation_gates_fired: [from salesforce-risk-taxonomy, or "none"]
missing_evidence: [what would improve the assessment]
assumptions: [explicit list]

Redaction rules

• Never request secrets, credentials, OAuth tokens, refresh tokens, session IDs, MFA seeds, customer PII.
• Sanitize org IDs, user IDs (replace with placeholders) before sharing in outputs.
• Field default values that appear to be real data must be flagged and excluded from output.

Privilege / data handling rules

• Works from schema-level exports only; does not require record-level data.
• If regulated-vertical indicators are present (Health Cloud, Financial Services Cloud ), escalate compliance findings to a qualified compliance specialist.
• Assessment output is internal; do not share org topology details externally.

Handoff rules

• Hands off to: salesforce-permission-model-review-skill (permission topology deep-dive), salesforce-flow-automation-review-skill (automation deep-dive), salesforce-metadata-review-skill (metadata type deep-dive), salesforce-integration-review-skill (integration deep-dive).
• If escalation gate fires: salesforce-case-capsule with escalation_required = true.
• Required handoff fields: matter_id, risk_register (summary), escalation_gates_fired, missing_evidence, assumptions.

Audit log fields

• matter_id, skill_id, skill_version, invoked_by, input_hash, evidence_quality, output_verdict, escalation_fired, timestamp

Stop conditions

• Export contains live credentials, session tokens, or record-level PII — stop and ask for sanitized version.
• Environment type is production and an escalation gate fires — stop and invoke salesforce-case-capsule before continuing.
• Export is from a regulated-vertical org and jurisdiction is unknown — flag and escalate.

Security notes

• Read-only static review; never requests live org access or API credentials.
• Sanitized inputs only; any input containing credentials must be refused.
• Risk register is advisory; remediation requires human authorization.
• Regulated-vertical findings must be escalated to qualified compliance counsel.