PBI Security Governance
SkillMonitoring & opsUse when creating or changing RLS roles, role filters, OLS permissions, perspectives, allow/deny/confirm policy rules, PII or numeric masking, or audit logging and evidence export, and when deciding whether a request needs security enforcement or only curation. For first-time guided setup of policies, masking, and guardrails, use mcp-engine-onboarding.
Available today. Use it from your connected AI after setup.
No other account needed.
Connect ahel once, and every AI you use reads what you have installed.
Then ask your AI: use the PBI Security Governance skill
What this skill tells your AI
The instructions your AI receives, as published by maxanatsko/mcp-engine-public in skills/mcp-engine-security-governance/SKILL.md and read by ahel’s review.
Security enforcement and governance through manage_security, manage_policy, and manage_audit. Keep curation separate from access control.
Start with the right category
- Perspectives (
create_perspective,update_perspective) are curation — they change what users see, not what they can access. - Roles, OLS, policies, masking, and audit are enforcement — fail closed by default unless the user explicitly changes that posture.
- Confirm the current model with
manage_model_connection{ "operation": "get_current" }and state the intended scope before modifying anything security-sensitive.
Branch by workflow
- RLS and OLS:
manage_security(create_role,set_role_filters,set_role_permissions) — read security-roles-guide first. For RLS, validate each affected role withrun_query{ "operation": "test_access", "query": "<small aggregate over the filtered data>", "spec": { "roles": ["<role>"] } }(bothqueryandspec.rolesare required). For OLS, exercise every changed table or column directly under the affected principals and verify each expected allow or denial; use targetedtest_accessqueries or amanage_testsols_validationmatrix. An unrelated aggregate is not OLS evidence. - Perspectives: perspectives-guide — when the user wants curated field visibility, not restricted access.
- Policy rules and packs:
manage_policy(status,evaluate,put,packs_applywithdry_run: truefirst) — policy-guide. Prefer deny rules overrequire_confirmfor high-risk operations;require_confirmis best-effort UX, not a portable control. - Masking: pii-masking-guide for sensitive-value redaction and masking behavior.
- Audit: audit-logging-guide for durable trails, export verification, and integrity checks. Treat auditability as part of correctness for write and governance flows.
Guardrails
- Confirm intent with the user before destructive or broad security changes.
- Never present a perspective as protection; say explicitly when a request needs RLS or OLS instead.
- Do not echo sensitive values, filter expressions over PII, or masked data in summaries, examples, or logs.
Report results
After security or governance work, report:
- Roles, permissions, policies, or masking rules created or changed.
test_accessorols_validationresults for each affected role, principal, and protected object.- Policy evaluation or dry-run outcomes before application.
- The enforcement posture after the change — what is now denied, confirmed, masked, or audited.
Signals
- GitHub stars
- 256
- Forks
- 65
- Last commit
- Aug 2026
Advanced
- Catalog kind
- skill
- Gateway key
mcp-engine-security-governance- Source
- github.com/maxanatsko/mcp-engine-public