io.github.Coalesce-Software-Inc/coalesce-transform
MCP server for managing Coalesce Transform workspaces, nodes, pipelines, and runs.
Using io.github.Coalesce-Software-Inc/coalesce-transform in Claude, Cursor, Gemini CLI, Cline, or Zed?
MCP tool contracts can change remotely with no version bump. The mcpindex gate pins each contract and HOLDs the call when it drifts—before your agent acts. Zero credentials. This is not the package install for this server itself (use Install this server for that).
Rewrites your MCP host config so each server launches behind the gate. Inspect first: curl -fsSL https://mcpindex.ai/install.sh | less
curl -fsSL https://mcpindex.ai/install.sh | shSemantic screen found no manipulation pattern in the description. Conformance probe not yet run.
mcpindex.integrity.descriptionpassINFOevidence“No malicious instructions found”via static_description
- - Semantic screen only - the deterministic conformance probe has not run on this server
- - Confidence is reported but not yet calibrated (v1)
- - Screen reads the tool description, not the live behavior
- - advisory
- - registry description only no input schema
- - screen model 8b
Semantic screen: an LLM judge reads the tool description for hidden instructions (status PARTIAL). A pass means the description is not lying, not that the tool is safe: a high-capability tool with an honest description still warrants caution. The deterministic conformance probe has not been run on this server yet, so the screen here is semantic-only. Posture: advisory. Confidences are reported but not yet calibrated (calibrated=false at v1). Full verdict history is not shown on this page.
Own this server? Screen its description →
A live verdict badge for your README or listing. It reflects the current screen, links back here, and updates when the verdict does.
[](https://mcpindex.ai/server/io-github-coalesce-software-inc-coalesce-transform)<a href="https://mcpindex.ai/server/io-github-coalesce-software-inc-coalesce-transform"><img src="https://mcpindex.ai/api/v1/badge/io-github-coalesce-software-inc-coalesce-transform" alt="mcpindex verdict" height="20" /></a>COALESCE_ACCESS_TOKENBearer token from the Coalesce Deploy tab. Optional when `~/.coa/config` provides a `token`.
COALESCE_PROFILESelects which `~/.coa/config` profile to load. Defaults to default.
COALESCE_BASE_URLRegion-specific base URL. Defaults to https://app.coalescesoftware.io (US).
COALESCE_ORG_IDFallback org ID for cancel-run. Also readable from `orgID` in the active ~/.coa/config profile.
COALESCE_REPO_PATHLocal repo root for repo-backed tools and pipeline planning. Also readable from `repoPath` in the active ~/.coa/config profile.
COALESCE_CACHE_DIRBase directory for the local data cache. When set, cache files are written here instead of the working directory. Also readable from `cacheDir` in the active ~/.coa/config profile.
COALESCE_MCP_AUTO_CACHE_MAX_BYTESJSON size threshold before auto-caching to disk. Defaults to 32768.
COALESCE_MCP_INLINE_RESPONSESForces large responses inline as text instead of auto-caching them to disk and returning a `resource_link` (and suppresses per-field cache resource links). The server already auto-enables this for clients known not to support MCP resources (e.g. Snowflake Cortex), which otherwise render the link as unreadable "binary data" and retry in a loop. Set `true` to force it on for an unrecognised client, or `false` to disable auto-detection. Leave unset for Claude/Cursor to keep large payloads out of context. Defaults to false (auto-enabled for resource-less clients).
COALESCE_MCP_LINEAGE_TTL_MSIn-memory lineage cache TTL in milliseconds. Defaults to 1800000.
COALESCE_MCP_MAX_REQUEST_BODY_BYTESMax outbound API request body size. Defaults to 524288.
COALESCE_MCP_REQUEST_TIMEOUT_MSDefault per-request timeout for Coalesce API calls. Raise on large workspaces where paginated reads exceed the default. Defaults to 60000.
COALESCE_MCP_DETAIL_FETCH_TIMEOUT_MSPer-page timeout for detail=true paginated fetches of large node lists. Defaults to 180000.
COALESCE_MCP_NODE_TYPE_RANKING_TIMEOUT_MSDeadline for the best-effort workspace-adoption node-type ranking lookup inside node creation. Advisory only — creation proceeds without a ranking suggestion if this trips. Defaults to 15000.
COALESCE_MCP_INVENTORY_CACHE_TTL_MSIn-memory TTL for the workspace node-type inventory. Set to `0` to disable. Defaults to 300000.
COALESCE_MCP_NODE_INDEX_CACHE_TTL_MSIn-memory TTL for the workspace node-index (id/name/type/location) cache used by SQL- and intent-ref resolution. Set to `0` to disable. Defaults to 300000.
COALESCE_MCP_NODE_DETAIL_CACHE_TTL_MSIn-memory TTL for the workspace node-detail (full body) cache used by review and predecessor lookups. Set to `0` to disable. Defaults to 300000.
COALESCE_MCP_READ_ONLYWhen `true`, hides all write/mutation tools during registration. Only read, list, search, cache, analyze, review, diagnose, and plan tools are exposed. Defaults to false.
COALESCE_MCP_SKILLS_DIRDirectory for customizable AI skill resources. When set, reads context resources from this directory and seeds defaults on first run. Users can augment or override any skill.
SNOWFLAKE_ACCOUNTSnowflake account identifier (e.g., `abc12345.us-east-1`). Required by the local `coa` CLI and `coa doctor`; not used by the MCP's REST run path. (required for run tools)
SNOWFLAKE_USERNAMESnowflake account username (required for run tools)
SNOWFLAKE_KEY_PAIR_KEYPath to PEM-encoded private key (required if SNOWFLAKE_PAT not set)
SNOWFLAKE_PATSnowflake Programmatic Access Token (alternative to key pair)
SNOWFLAKE_KEY_PAIR_PASSPassphrase for encrypted keys
SNOWFLAKE_WAREHOUSESnowflake compute warehouse (required for run tools)
SNOWFLAKE_ROLESnowflake user role (required for run tools)
Resolve .hood names on Robinhood Chain — forward/reverse, text records, availability & pricing.
AI-powered trading strategy development: backtesting, market data, and portfolio analysis
Feature flagging and A/B testing platform with AI-first experimentation workflows.