SLM Agent by ScaleDown vs Cost Guard MCP
SLM Agent by ScaleDown is a plugin bundle for AI coding tools published by scaledown-team. Marketplaces we read that list it: 1 (Claude Code: claude-plugins-community). Cost Guard MCP is an MCP server published by mcpsmiths. Marketplaces we read that list it: 2 (Official MCP Registry, PyPI (MCP servers)). As of our marketplace pass on 2026-10-28. This pair was picked by a reader, not by our comparison rule, so search engines are asked not to index the page.
Where each is published
| Marketplace |  SLM Agent by ScaleDown |  Cost Guard MCP |
|---|---|---|
| Claude Code: claude-plugins-community | Listed Version: 0.1.7 Forks (GitHub): 8 Stars (GitHub): 38 Not marked · First seen |
Not listed |
| Official MCP Registry | Not listed | Listed Version: 0.6.0 Downloads/month (PyPI): 777 Downloads/week (PyPI): 365 Forks (GitHub): 0 Stars (GitHub): 1 Not marked · First seen |
| PyPI (MCP servers) | Not listed | Listed Version: 0.6.0 Downloads/month (PyPI): 777 Downloads/week (PyPI): 365 Forks (GitHub): 0 Stars (GitHub): 1 Not marked · First seen |
Each figure is the marketplace's own and is never added up across marketplaces. A marketplace that publishes no figure shows none: no number is not zero.
The facts side by side
| Fact | SLM Agent by ScaleDown | Cost Guard MCP |
|---|---|---|
| Marketplaces that list it (differs) | 1 (worse) | 2 (better) |
| Stars (GitHub) (differs) | 38 (better) | 1 (worse) |
| Forks (GitHub) (differs) | 8 (better) | 0 (worse) |
| Last release (differs) | Not published | 2026-10-20 |
| Where it runs (differs) | On your machine (@scaledown/migration-agent) | On your machine (cost-guard-mcp) |
| Transport (differs) | Not stated | stdio |
| Tools it declares | Not published | Not published |
| Environment variables it asks for (differs) | Not published | 19, secrets among them: 6 |
| Parts of the bundle (differs) | 3 | Not a bundle |
| Licence | MIT | MIT |
| Repository (differs) | github.com/scaledown-team/slm_agent | github.com/mcpsmiths/cost-guard-mcp |
| Changes in the last 7 days | 0 | 0 |
| First seen (differs) | 2026-10-12 | 2026-10-11 |
Lines where the two differ have a tinted background and bold values. Where the data itself says which value is better (more marketplaces, a larger figure of the same marketplace, a fresher release, a verified mark on a marketplace that lists both), the better one is also marked green with ▲ and the worse one red with ▼. Where it runs, the transport, the tools, the settings, hooks, the bundle's parts, the licence and the week's changes are marked only as different: more or fewer is not better in itself.
What each says it does
SLM Agent by ScaleDown
SLM Agent scans your codebase for AI API calls and generates a structured migration plan showing exactly where you can swap frontier LLM calls (Opus) for Small Language Models API calls like Haiku or Sonnet. It performs deep cross-file analysis: tracing prompts to their construction site, following callers across files, and classifying each call by what it actually does. Each finding includes a complexity score, confidence level, estimated savings, and for multi-task calls, a step-by-step decomposition. The output is a saved report.md in your project with a full API reference, so you can migrate. No source code or codebase data is sent to any external server (except Claude) during analysis.
Cost Guard MCP
Pre-flight query cost and result-size guardrails for AI agents on BigQuery, Snowflake, Databricks
How to run each
SLM Agent by ScaleDown As a Claude Code plugin
/plugin marketplace add anthropics/claude-plugins-community
/plugin install slm-agent@claude-communityCost Guard MCP As a local MCP server: PyPI package
claude mcp add --env "DATABRICKS_CLIENT_SECRET=<YOUR_DATABRICKS_CLIENT_SECRET>" --env "DATABRICKS_TOKEN=<YOUR_DATABRICKS_TOKEN>" --env "GOOGLE_APPLICATION_CREDENTIALS=<YOUR_GOOGLE_APPLICATION_CREDENTIALS>" --env "SNOWFLAKE_PASSWORD=<YOUR_SNOWFLAKE_PASSWORD>" --env "SNOWFLAKE_PRIVATE_KEY_PASSPHRASE=<YOUR_SNOWFLAKE_PRIVATE_KEY_PASSPHRASE>" --env "SNOWFLAKE_PRIVATE_KEY_PATH=<YOUR_SNOWFLAKE_PRIVATE_KEY_PATH>" --transport stdio cost-guard-mcp -- uvx cost-guard-mcpRequires: DATABRICKS_CLIENT_SECRET (secret), DATABRICKS_TOKEN (secret), GOOGLE_APPLICATION_CREDENTIALS (secret), SNOWFLAKE_PASSWORD (secret), SNOWFLAKE_PRIVATE_KEY_PASSPHRASE (secret), SNOWFLAKE_PRIVATE_KEY_PATH (secret)
Commands built only from what the marketplaces declare; the full set for every client is on each card.
Tools each declares
Neither one's marketplaces publish a list of tools, so there is nothing to compare.
Settings each asks for
The marketplaces of SLM Agent by ScaleDown publish no settings it asks for, so we do not compare settings.
What it can do: SLM Agent by ScaleDown · What it can do: Cost Guard MCP · Compare two other plugins
Data comes from the public marketplace file https://github.com/anthropics/claude-plugins-community/blob/HEAD/.claude-plugin/marketplace.json. Data comes from the Official MCP Registry (registry.modelcontextprotocol.io), CC0 1.0. Package data comes from the Python Package Index (pypi.org). Figures come from our own regular passes over the marketplaces, last read on . We publish no contact details for plugin publishers.