using-sap-superpowers — independently scanned and version-tracked by SaferSkills.
SaferSkills independently audited using-sap-superpowers (Agent Skill) and scored it 100/100 (green). The audit ran 55 deterministic rules across Security, Supply Chain, Maintenance, Transparency, and Community; it found 0 high-severity and 0 lower-severity findings. The full rule-by-rule trace and per-finding evidence are below. Free, methodology-open.
Findings & checks · 0 flagged
Every scanned point with the score it earned and what moved between them.
First recorded scan — no prior version to compare against.
The primary manifest — the file an agent reads to learn what this artifact does.
This skill enforces correct use of the SAP Superpowers system — routing every request through the right skill, following workflow chains, leveraging the knowledge layer, and delegating to specialized agents.
verification-before-completion if there is any doubt. Unverified work is unfinished work.knowledge/ files. Do not rely on training data alone. If the knowledge layer lacks what you need, state the gap explicitly.| Category | Skills | Activates When |
|---|---|---|
| Meta (3) | using-sap-superpowers, verification-before-completion, writing-sap-skills | Every session (auto-injected); when creating new skills; when completing any task |
| Consulting (8) | sap-project-kickoff, sap-brainstorming, sap-estimation, sap-fit-gap-analysis, sap-solution-architecture, sap-value-advisory, sap-change-management, sap-process-design | Project planning, scoping, estimation, architecture, organizational change |
| Development (5) | sap-development-workflow, sap-code-review, sap-troubleshooting, sap-code-generation, sap-performance-tuning | Writing code, reviewing code, debugging, generating ABAP/CDS/RAP, optimizing performance |
| Delivery (5) | sap-testing-strategy, sap-data-migration, sap-cutover-planning, sap-go-live-readiness, sap-hypercare | Test planning, data loads, cutover prep, go-live decisions, post-go-live support |
| Module Reference (20) | abap-cloud, btp, integration-suite, system-admin, security-grc, FI, CO, MM, SD, PP, PM, SF, analytics, EWM, TM, Ariba, Fieldglass, Concur, QM, Joule | Module-specific questions, configuration, transactions, best practices |
| Strategic (5) | s4hana-migration, activate-methodology, rise-licensing, clean-core-strategy, joule | Migration planning, methodology questions, licensing, clean core adoption |
Skills chain together following SAP Activate phases. Each skill explicitly names the next skill to invoke. Never treat a skill as an island.
DISCOVER : sap-project-kickoff --> sap-brainstorming
PREPARE : sap-estimation --> sap-solution-architecture --> sap-change-management
EXPLORE : sap-fit-gap-analysis --> sap-process-design --> [module skills]
REALIZE : sap-development-workflow --> sap-code-review --> sap-testing-strategy
DEPLOY : sap-data-migration --> sap-cutover-planning --> sap-go-live-readiness
RUN : sap-hypercare --> sap-value-advisoryHow chaining works:
The knowledge/ directory contains curated, up-to-date SAP reference data:
| Knowledge File | Contains | Use When |
|---|---|---|
sap-notes-index.md | Critical SAP Notes by area | Referencing known issues, patches, corrections |
transaction-codes.md | Comprehensive tcode reference | User asks about a tcode or you need to recommend one |
released-apis.md | Released API catalog with replacements | ABAP Cloud development, clean core checks |
simplification-items.md | S/4HANA simplification items | Migration analysis, compatibility checks |
glossary.md | SAP terminology with context | Ensuring correct term usage |
best-practices/*.md | Module-specific best practices | Any module-specific recommendation |
patterns/*.md | Reusable solution patterns | Architecture, development patterns |
Rules for knowledge use:
Dispatch specialized agents for focused, parallel work:
| Agent | Specialization | Dispatch When |
|---|---|---|
sap-reviewer | Code review findings | Code review skill needs analysis |
sap-estimator | Effort decomposition | Estimation requires deep breakdown |
sap-migration-analyzer | S/4HANA readiness | Migration compatibility checks |
sap-test-designer | Test case generation | Testing strategy needs test cases |
sap-value-calculator | ROI/TCO modeling | Business case needs financial analysis |
sap-security-auditor | Auth and security patterns | Security review needed |
sap-architect | Solution architecture review | Architecture validation |
sap-data-analyst | Data migration analysis | Data quality, mapping, transformation |
sap-doc-generator | Documentation generation | Deliverable creation at scale |
sap-process-modeler | Process flow design | Business process mapping |
Delegation rules:
When a user request arrives, route it through this decision tree:
writing-sap-skills| Agent Will Try To... | Why It Seems Reasonable | Why It Fails | Counter |
|---|---|---|---|
| Answer directly without invoking a skill | "It's a simple SAP question, I already know the answer" | Uninstrumented answers have no iron laws, no verification, no deliverable. Quality is uncontrolled. | Route every SAP request to a skill. Simple questions still get the skill's structure. |
| Skip the workflow chain | "The user only asked about estimation, not architecture" | Estimation without architecture context produces numbers disconnected from reality. The chain exists for a reason. | Always inform the user about the next skill in the chain. |
| Ignore the knowledge layer | "I have enough training data about SAP" | Training data can be outdated — BSEG vs. ACDOCA, deprecated BAPIs, changed tcodes. The knowledge layer is curated. | Check knowledge files first for any factual SAP claim. |
| Handle everything without agents | "I can do the code review and security check myself" | Attempting everything serially produces shallow analysis. Agents go deep on focused tasks. | Dispatch agents for tasks that benefit from focused specialization. |
| Generate guidance instead of deliverables | "I explained the process thoroughly" | Explanations are not deliverables. The client cannot use an explanation in a steering committee. | Every process skill must produce a file. No file = not done. |
| Merge multiple skills into one response | "I'll cover estimation and architecture together to save time" | Merged skills skip individual verification gates. Each skill has its own iron laws for a reason. | Execute skills sequentially. Each must complete verification before the next begins. |
Watch for these phrases in your own reasoning — each signals a system bypass:
If you catch yourself thinking any of these: STOP. Re-read the routing logic. Find the right skill. Follow it completely.
<HARD-GATE> Before responding to ANY SAP-related request:
Proceeding without skill identification is a system violation. </HARD-GATE>
This skill is complete ONLY when ALL of the following are true:
verification-before-completion was satisfied for each skillEvidence required: Named skill(s) invoked, deliverable files produced, knowledge sources cited, agent dispatches logged, chain status communicated.
If any verification item is not met, the session is NOT complete. Do not claim completion.
This skill is always active. It does not chain to a single next skill — it routes to the appropriate skill based on the user's request. After routing and completing the target skill, that skill's "Next Skill" section governs the chain.
verification-before-completion — Invoke before claiming ANY task is donewriting-sap-skills — Invoke when creating or modifying skillsknowledge/ layer — Referenced for all factual SAP claimsagents/ — Dispatched for focused subtasks per the Agent Delegation table~30 seconds. Free. No account. Every finding cites a rule and a line of evidence.