Loading ToolBench

Reading the current corpus…

mcp-gong MCP server · D · ToolBench

← Browse

mcp-gong

github.com/pipeworx-io/mcp-gong

Gong MCP — wraps the Gong API v2 (OAuth)

2026-07-28+5 tools

55/100

Definition Quality71

Protocol Readiness35

Supportability43

Spec posture

Inferred effective spec 2026-07-28+·current baseline 2026-07-28

Static source inference · medium confidence · detected: Dynamic Client Registration

Deprecated protocol patterns detected

Summary

Gong MCP has well-structured tool definitions with clear naming (verb_noun pattern) and reasonable descriptions (avg 120 chars). All 5 tools have input schemas with type declarations. However, parameter descriptions are sparse or missing entirely, most parameters lack guidance on format, constraints, or valid values. Output schemas are completely undocumented, forcing LLMs to guess what fields to expect. Error handling is minimal (generic 'Gong API error' messages). No pagination guidance in list_calls despite cursor support. Token injection via _context is non-standard and undocumented.

Tools (5)

gong_get_callread onlyauthsource verified73/100

Get details for a specific Gong call by its ID, including participants, duration, and metadata.

gong_get_transcriptread onlyauthsource verified72/100

Get the transcript for a specific Gong call. Returns the full conversation transcript.

gong_list_callsread onlyauthsource verified72/100

List recorded calls from Gong. Optionally filter by date range. Supports cursor-based pagination.

gong_list_usersread onlyauthsource verified65/100

List all users in the Gong workspace.

gong_search_callsread onlyauthsource verified72/100

Search Gong calls by keyword. Returns calls that match the search text.

Issues (5)

CRITICALtool ↗

Output schemas completely undocumented. LLMs cannot predict what fields gong_list_calls, gong_get_call, gong_get_transcript, gong_list_users, or gong_search_calls return. This forces guessing and breaks downstream tool chaining.

gong_list_callsgong_get_callgong_get_transcriptgong_list_usersgong_search_calls

HIGHtool-description ↗

Parameter descriptions missing or incomplete. 'fromDateTime' and 'toDateTime' lack format guidance (ISO 8601 is mentioned in description but not in param docs). 'cursor' has no explanation of pagination behavior. 'keyword' has no length/format constraints.

gong_list_callsgong_search_calls

HIGHsecret-injection ↗

Token injection via _context is non-standard and undocumented. The getToken() helper expects args._context.gong.accessToken but this pattern is not explained in tool descriptions or server documentation. LLMs cannot infer this requirement.

gong_list_callsgong_get_callgong_get_transcript

Recommendations

Document output schemas for all 5 tools. Example: gong_list_calls should return {calls: [{callId, participants, duration, ...}], cursor?, total?}. This enables LLMs to chain tools and extract relevant fields.
Add parameter descriptions with format/constraint details. E.g., 'fromDateTime: ISO 8601 timestamp (e.g., 2024-01-01T00:00:00Z). Required if filtering by date range.' for gong_list_calls.
Replace _context token injection with standard MCP secret handling. Document how OAuth tokens are provisioned (via environment, config, or MCP initialize handshake) so LLMs understand the prerequisite.
Enhance error messages with recovery guidance. E.g., 'Gong API 401: Invalid token. Ensure OAuth credentials are configured.' or 'Gong API 404: Call not found. Verify callId with gong_list_calls first.'
Add pagination guidance to gong_list_calls: 'Returns up to 50 calls per request. Use cursor from response to fetch next page. Stop when cursor is null or empty.'

Score history

Overall score trend

First recorded score · v2 rubric

55/100

Scored	Grade	Overall	Spec posture	Rubric
2026-09-23	D	55	2026-07-28+	v2