Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx agentmods add skills/vigneshbarani24/sap-superpowers/code-reviewnpx skills add vigneshbarani24/sap-superpowers --skill code-reviewgit clone --depth 1 https://github.com/vigneshbarani24/sap-superpowersWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/vigneshbarani24/sap-superpowers/code-review)<a href="https://agentmods.dev/skills/vigneshbarani24/sap-superpowers/code-review"><img src="https://agentmods.dev/badge/skills/vigneshbarani24/sap-superpowers/code-review.svg" alt="Measured on agentmods" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00036 | $0.02728 |
| Opus 5 | $0.00018 | $0.01364 |
| Sonnet 5 | $0.00007 | $0.00546 |
| Haiku 4.5 | $0.00004 | $0.00273 |
Grade A, and why
code-review scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 5d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 128 lines — stays where its author put it; the contents beside it link to each section on GitHub.
SAP Code Review
This skill enforces zero-tolerance code quality gates — it makes approving dangerous, non-performant, or insecure SAP code structurally impossible.
Iron Laws
- NEVER APPROVE SELECT IN LOOPS. A SELECT statement inside a LOOP...ENDLOOP, DO...ENDDO, or WHILE...ENDWHILE is an automatic rejection. No exceptions. Not even for "small tables." One SELECT in a loop that runs against a large dataset is a production system outage.
- NEVER APPROVE CODE WITHOUT AUTHORITY CHECKS. Any code that reads, writes, or executes against business data must have explicit AUTHORITY-CHECK statements. "The calling program already checks authority" is not acceptable — each component enforces its own authorization.
- NEVER APPROVE MODIFICATIONS TO SAP STANDARD. No direct modification to SAP-delivered objects. Enhancements must use released BAdIs, User Exits, Enhancement Spots, or CDS extensions. SAP standard modifications are overwritten at every support package upgrade.
- NEVER APPROVE WITHOUT ATC CLEAN. Code that has not been run through ABAP Test Cockpit (ATC) with no open Priority 1 or Priority 2 findings cannot be approved. ATC findings are not suggestions — they are known defects.
- NEVER APPROVE CODE THAT ACCESSES DEPRECATED OR LOCKED TABLES DIRECTLY. In S/4HANA, tables like BSEG, VBAP, MARA have compatibility views or released CDS entities. Direct table access that bypasses the virtual data model creates upgrade incompatibilities.
Rationalization Table
| Agent Will Try To... | Why It Seems Reasonable | Why It Fails | Counter |
|---|---|---|---|
| Approve a SELECT in a loop "because the table is small" | "There are only 100 records in that table now" | Tables grow. A table with 100 records today will have 1,000,000 in two years. The code will still be in production. | Iron Law 1. No SELECT in loops, regardless of current table size. Refactor to SELECT...FOR ALL ENTRIES or JOIN. |
| Skip authority check review because "the Fiori app controls access" | "The tile and role assignment handle authorization" | Direct RFC calls, background jobs, and API access all bypass Fiori tile authorization. Every executable unit must check authority independently. | Iron Law 2. Show me the AUTHORITY-CHECK statement. If it doesn't exist, the review fails. |
| Approve a modification because "it's the only way" | "There's no BAdI for this, modification is the only option" | There is almost always a released API, Enhancement Framework entry, or alternative approach. "Only way" usually means "fastest way I found." | Checklist Step 5: Document the three alternatives investigated before modification is considered. Modification requires architect sign-off. |
| Treat ATC findings as optional | "These are just warnings, not real errors" | ATC Priority 2 findings include security vulnerabilities, performance anti-patterns, and clean core violations. They are categorized warnings for a reason. | Iron Law 4. ATC clean is a gate, not a guideline. P1 and P2 findings must be resolved or formally suppressed with justification. |
| Approve code using BSEG directly in S/4HANA | "BSEG still exists in S/4HANA, it works" | BSEG in S/4HANA is a compatibility view. Direct access is inefficient and will break at clean core enforcement. Use I_JournalEntryItem or equivalent released CDS. | Iron Law 5. Direct access to compatibility views is a clean core violation. Use released APIs. |
| Skip performance review for "simple" programs | "It's just a small report" | "Small reports" run in batch. Batch programs process full table sets. A report that takes 5 seconds per record × 1M records = 6 days of runtime. | Checklist Step 3: Performance review applies to every program, regardless of perceived complexity. |
| Approve hardcoded values without constants | "The value is obvious from context" | Hardcoded values become silent bugs when configuration changes. Magic numbers and strings are the root cause of countless production defects. | Checklist Step 4: Constants, not hardcoded literals. Every unexplained literal is a review finding. |
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 5d ago First seen · 128 lines · 36 tokens per session scan A 16e3204a839f
code-review is a skill published in the GitHub repository vigneshbarani24/sap-superpowers (9 stars, last pushed 13d ago), licensed MIT. It adds 36 tokens to every session and 2,728 once invoked, about $0.0002 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other skills, from other repositories
sap-expert
Expert in SAP ERP systems, ABAP programming, SAP HANA, S/4HANA, Fiori applications, and SAP integration patterns including OData, RFC, and IDoc. Use when the user mentions ERP, enterprise, business apps, ABAP, HANA, or S/4HANA, or when the task involves SAP Ecosystem, ABAP Development, Integration Technologies, or…
migrate-from-openclaw
Migrate from OpenClaw to NanoClaw v2. Detects an existing OpenClaw installation, extracts identity, channel credentials, scheduled tasks, and other config, then guides interactive migration. Triggers on "migrate from openclaw", "openclaw migration", "import from openclaw".
add-dial-tool
Give chosen NanoClaw agents a real phone number as a container tool — the dial CLI baked into the agent image plus OneCLI credential injection for api.getdial.ai, scoped per agent, so the agents you pick can send SMS, place AI voice calls, and receive verification codes from inside the sandbox. Independent of the Dial…
slack-agent-flow
Let an existing Slack agent create new agents that arrive as their own Slack bots — provisioned app, operator DM, and a shared three-way room, hot-started without a host restart.
add-ollama-tool
Add Ollama MCP server so the container agent can call local models and optionally manage the Ollama model library.
add-signal
Add Signal channel integration via signal-cli device-link. Native adapter — no Chat SDK bridge.