Loading ToolBench
MCP server for Salesforce B2C Commerce Cloud development assistance including logs, debugging, and development tools
The SFCC Dev MCP server has 28 tools with reasonably comprehensive descriptions and schemas, but exhibits several quality gaps that prevent a higher score. Naming is verb-based and clear (sync_, disable_, get_, search_, list_, generate_, activate_). Descriptions are generally present and moderately detailed (averaging 150-250 chars), meeting baseline minimums. However, parameter descriptions are inconsistent, many core parameters lack detail, and several parameter types lack explicit enumeration constraints. Output schemas are not visibly documented in the source code provided, forcing inference. Error handling and recovery guidance are minimal. Tool composition is reasonable but some tools combine multiple concerns (e.g., sync_agent_instructions handles both AGENTS.md and skills with multiple merge strategies). Security considerations around file writes are not evident in descriptions. Overall, the server is functional and above-average for community tools, but falls short of production-grade quality expected at 70+.
Activate a code version (deactivates current). Use for code-switch fixes, SCAPI endpoint issues, or deployment conflicts. Only inactive versions can be activated.
Creates or updates mcp-dev.json in the project root with {"disableAgentSync": true} to permanently disable agent sync suggestions. Call this when the user declines to install AGENTS.md and skills, so they will not be prompted again.
Generate a complete SFCC cartridge with proper directory structure, configurations, and boilerplate files. Creates files directly in the target directory.
List all code versions on the SFCC instance. Use for deployment management, identifying active version, or preparing code-switch fixes.
Get all ISML element categories with descriptions and counts.
Get detailed documentation for an ISML element including syntax, attributes, examples, and use cases.
Output schemas not documented in source. Return types and field structures are not visible in the tool definitions, forcing LLM inference and risking context loss or incorrect field selection.
Missing parameter descriptions for common parameters. 'date' parameter in log tools lacks detail about format validation, acceptable ranges, or examples. 'limit' parameters lack min/max constraints.
Inferred effective spec: 2025-06-18+.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-22 | C | 66 | 2025-06-18+ | v2 |
| 2026-03-09 | F | 0 | - | v1 |
Get ISML elements filtered by category (control-flow, output, includes, scripting, cache, decorators, special, payment, analytics).
Get execution summary for a job: timing, status, error counts, and step info. Use for monitoring job health and performance.
Get job log entries by level (error/warn/info/debug/all). Unlike standard logs, job logs combine all levels in one file.
Get latest debug messages. Use for step-by-step tracing, variable inspection, or when error logs lack context.
Get latest SFCC error log entries. Use for debugging crashes, exceptions, API failures, or when code throws errors.
Get latest info messages. Use to verify operations completed, track business logic execution, or monitor normal flow.
Get recent job log files from /Logs/jobs/. Job logs contain all log levels in single files. Use for debugging custom job steps.
Get latest warning messages. Use to identify deprecations, potential issues, or suboptimal configurations.
Read full contents of a specific log file. Use for detailed analysis when you need complete error traces or full context.
Get the full raw documentation for an SFCC class including examples and detailed descriptions. Use when get_sfcc_class_info lacks sufficient detail.
Get detailed information about an SFCC class including properties, methods, constants, and inheritance. Essential for understanding dw.* APIs when building controllers, scripts, templates, or REST APIs. Supports filtering by section and search within class members.
List all ISML template elements with summaries. Includes control flow (isif, isloop), output (isprint), includes (isinclude, iscomponent), scripting (isscript), and caching (iscache).
List available log files with sizes and modification dates. Use to discover available log data or check log retention.
List all available SFCC dw.* classes organized by namespace. Use to explore the full API surface or find classes in a specific domain.
Search ISML elements by purpose, attribute name, or use case. Returns relevance-scored results with preview snippets.
Search job logs for patterns, error messages, or custom logging from job steps. Essential for debugging custom job code.
Find job log files by job name (partial match supported). Use to locate logs for a specific job.
Search logs for specific patterns, error messages, order numbers, user IDs, or custom identifiers. Essential for tracking specific transactions.
Find SFCC classes by partial name or keyword. Use when you don't know the exact class name (e.g., 'catalog' finds dw.catalog.*, 'price' finds pricing-related classes).
Find methods across all dw.* classes by name. Use when you know a method exists but not which class contains it.
Get a health overview of all log activity with counts and key issues. Use as first step when investigating problems or for daily health checks.
ONLY FOR SALESFORCE B2C COMMERCE (SFCC) PROJECTS. IMPORTANT: Before calling this tool, ALWAYS ASK THE USER if they want to install AGENTS.md and bundled SFCC skills into their project. If the user declines, do NOT call this tool - instead call disable_agent_sync to permanently disable this suggestion. This tool installs/merges AGENTS.md and bundled SFCC skills into the project (or a temp dir for manual merge). Supports dry-run planning, append/replace/skip merge strategies, and selecting skills directory (.github/skills, .agents/skills, .claude/skills, .agent/skills, .cursor/skills, and user-home defaults like ~/.copilot/skills).
Minimal error handling guidance. Descriptions do not indicate what the LLM should do if a search returns no results, a file is not found, or a code version is already active.
sync_agent_instructions tool combines multiple concerns (AGENTS.md + skills, with append/replace/skip merge strategies) and writes to multiple possible locations (project/user/temp). Complex parameter interdependencies are not documented.
File write operations (sync_agent_instructions, generate_cartridge_structure, disable_agent_sync) lack safety warnings or confirmation steps. No mention of backup, rollback, or dry-run consequences.
Pagination support is inferred but not formally documented. Tools like search_logs accept 'limit' but no 'offset', 'cursor', or 'page' parameters, and return structures do not indicate total counts or next-page availability.
activate_code_version tool has destructive intent (activates a version, deactivates current) but lacks confirmation or dry-run support, and description does not warn of irreversibility.