[mcp-analysis] MCP Structural Analysis - 2026-08-13

[github-actions[bot]]

Analyzed 10 GitHub MCP tool responses across 9 toolsets on 2026-08-13 (day 2 of trending). Average usefulness: 3.6/5. Best: github_context (5/5, zero-overhead identity block). Worst: get_teams (1/5, blocked by an ungranted permission for the second consecutive day). Notable finding: get_file_contents on README.md jumped from ~4,500 to ~11,300 tokens day-over-day as the file grew to 42.8KB, driven mostly by an embedded 280+ entry community-contributor link list rather than the useful content.

Full Structural Analysis Report

Executive Summary

Metric	Value
Tools Analyzed	10
Total Tokens (Today)	13,825
Average Usefulness Rating	3.6/5
Best Rated Tool	github_context: 5/5
Worst Rated Tool	get_teams: 1/5

Usefulness Ratings for Agentic Work

Tool	Toolset	Rating	Assessment
github_context	workflow_context	⭐⭐⭐⭐⭐	Excellent — zero-MCP-overhead identity block
list_pull_requests	pull_requests	⭐⭐⭐⭐	Good, minimal field-filtered response
list_workflows (mcpscripts wrapper)	actions	⭐⭐⭐⭐	Good — wrapper correctly honors perPage, unlike the built-in tool
list_code_scanning_alerts	code_security	⭐⭐⭐⭐	Good, clean empty-result handling
list_discussions	discussions	⭐⭐⭐⭐	Good — cursor pagination and category/author present
get_label	labels	⭐⭐⭐⭐	Good — single-label lookup much leaner than bulk list_label
search_code	search	⭐⭐⭐⭐	Good, well-scoped with perPage=1
get_file_contents	repos	⭐⭐⭐	Adequate — core content clean but bloated by contributor list + icons
list_issues	issues	⭐⭐⭐	Adequate — integrity-policy redaction obscures true result count
get_teams	context	⭐	Poor — blocked by permission not granted, zero data two days running

Schema Analysis

Tool	Type	Depth	Key Fields	Notes
github_context	markdown	1	actor, repository, workspace, workflow-run-id	Prompt-injected, no MCP call
get_teams	error	0	—	Permission-gated, no schema to analyze
get_file_contents	object	4	content, resource.text, resource.uri, _meta.serverInfo	Deeply nested wrapper around raw file text
list_issues	array	2	number, title, state	Field-filtered; one item redacted by integrity policy
list_pull_requests	array	2	number, title, state	Field-filtered, flat
list_workflows	object	3	total_count, workflows[].id/name/path/state/html_url	Clean object array, correct pagination via wrapper
list_code_scanning_alerts	array	1	—	Empty array, no nesting
list_discussions	object	3	discussions[].category.name, title, user.login, pageInfo, totalCount	Cursor-based pagination present
get_label	object	1	color, description, id, name	Flat single-entity object
search_code	object	3	total_count, items[].name, items[].path	Clean, match locations omitted via field filter

Response Size Analysis

Toolset	Avg Tokens	Tools Tested
repos	11,300	1
discussions	800	1
search	660	1
labels	650	1
actions	180	1
pull_requests	30	1
issues	90	1
workflow_context	90	1
code_security	10	1
context	15	1

Tool-by-Tool Analysis

Tool	Toolset	Tokens	Schema	Rating	Notes
github_context	workflow_context	90	markdown	5	Stable day-over-day, zero overhead
get_teams	context	15	error	1	Ungranted permission blocked the call again
get_file_contents	repos	11,300	object	3	README.md grew to 42.8KB (confirmed via wc -c); bulk is a 280+ entry community list plus ~2.5KB duplicated base64 icons
list_issues	issues	90	array	3	Integrity policy silently redacted 1 matching issue
list_pull_requests	pull_requests	30	array	4	Clean, minimal, consistent with prior day
list_workflows	actions	180	object	4	mcpscripts wrapper correctly honored perPage=1 out of 360 workflows — confirms this is the fix for the built-in tool's camelCase/snake_case pagination bug
list_code_scanning_alerts	code_security	10	array	4	Correctly scoped state=open, severity=high; empty result
list_discussions	discussions	800	object	4	totalCount now 9,496 (up from prior day); fixed icon overhead is main cost
get_label	labels	650	object	4	Single-label lookup far leaner than bulk list_label (868 labels, 3,250 tokens)
search_code	search	660	object	4	Well-scoped; text_matches omitted via field filter for smaller payload

30-Day Trend Summary

Metric	Value
Data Points	20
Days Tracked	2 (2026-08-12, 2026-08-13)
Average Daily Tokens	~9,860
Average Rating Trend	Stable (3.6 today vs. comparable distribution prior day), except get_file_contents size regressed sharply

Recommendations

• High-value tools (rating 4-5): github_context, list_pull_requests, list_workflows (via mcpscripts wrapper), list_code_scanning_alerts, list_discussions, get_label, search_code.
• Tools needing improvement: get_teams (permission-gating prevents any use two runs in a row — needs the permission pre-granted or an alternative context tool); list_issues (integrity-policy redaction should surface a machine-readable reason rather than silently dropping items); built-in list_workflows/list_label (ignore perPage/have no pagination — use the mcpscripts wrapper for list_workflows and prefer single-item get_label over bulk list_label when possible).
• Context-efficient tools (low tokens, high rating): list_pull_requests (30 tokens, 4/5), list_code_scanning_alerts (10 tokens, 4/5), list_workflows via wrapper (180 tokens, 4/5).
• Context-heavy tools: get_file_contents on README.md (11,300 tokens) — consider a ranged/section read option for large files with embedded generated content like contributor lists.

Visualizations

Response Size by Toolset

Usefulness Ratings

Daily Token Trend

Size vs Usefulness

Generated by 🔍 GitHub MCP Structural Analysis · agent · 116.5 AIC · ⌖ 45.6 AIC · ⊞ 11.8K · ◷

• expires on Aug 14, 2026, 4:18 AM UTC-08:00

[github-actions[bot]]

This discussion has been marked as outdated by GitHub MCP Structural Analysis.

A newer discussion is available at Discussion #52716.