[mcp-analysis] MCP Structural Analysis - 2026-08-05

[github-actions[bot]]

MCP Structural Analysis — 2026-08-05

10 GitHub MCP tools tested across 10 toolsets. Total: 7,525 tokens, average usefulness 3.6/5. Best: github_context and list_pull_requests (5/5). Worst: get_teams (1/5, blocked by permissions again).

Notable finding: the mcpscripts list_workflows wrapper correctly honored perPage=1 today (1 of 351 workflows returned, 180 tokens) — a clear improvement over the raw GitHub MCP list_workflows tool tested 2026-08-04, which ignored per_page and returned 30 of 347 workflows (2,200 tokens).

Full Structural Analysis Report

Executive Summary

Metric	Value
Tools Analyzed	10
Total Tokens (Today)	7,525
Average Usefulness Rating	3.6/5
Best Rated Tool	github_context / list_pull_requests: 5/5
Worst Rated Tool	get_teams: 1/5

Usefulness Ratings for Agentic Work

Tool	Toolset	Rating	Assessment
github_context	workflow_context	⭐⭐⭐⭐⭐	Excellent — zero-cost identity, always available
list_pull_requests	pull_requests	⭐⭐⭐⭐⭐	Excellent — compact, exactly matches perPage/fields
list_workflows (mcpscripts)	actions	⭐⭐⭐⭐⭐	Excellent — perPage correctly honored this run
list_code_scanning_alerts	code_security	⭐⭐⭐⭐	Good — clean empty-array response
list_discussions	discussions	⭐⭐⭐⭐	Good — compact, includes cursor + related entities
search_code	search	⭐⭐⭐⭐	Good — succeeded this run, useful text_matches
list_issues	issues	⭐⭐⭐	Adequate — result redacted by integrity policy this run
list_label	labels	⭐⭐⭐	Adequate — no perPage/filter, always returns full set
get_file_contents	repos	⭐⭐	Limited — no excerpt support, README dominated by boilerplate
get_teams	context	⭐	Poor — blocked by permissions two runs in a row

Schema Analysis

Tool	Type	Depth	Key Fields	Notes
github_context	markdown	1	actor, repository, workspace, workflow-run-id	Prompt-injected, no MCP call
get_file_contents	object	2	content, resource.text, resource.uri	Whole-file only, no range read
list_issues	array	1	(redacted)	Integrity policy stripped the one candidate result
list_pull_requests	array	1	number, title, state, html_url	Clean, flat
list_workflows	object	2	total_count, per_page, workflows[].id/name/path/state	perPage honored via mcpscripts wrapper
list_code_scanning_alerts	array	1	(empty)	No alerts open
list_discussions	object	3	discussions[].number/title/category/user, pageInfo, totalCount	Cursor pagination + embedded entities
list_label	object	2	labels[].id/name/color/description, totalCount	No pagination control
search_code	object	3	total_count, items[].name/path/repository/text_matches	Inline match fragments

Response Size Analysis

Toolset	Avg Tokens (30-day)	Tools Tested
repos	4,500	2
labels	2,550	2
actions	1,190	2
discussions	140	2
search	125	2
workflow_context	80	2
issues	60	2
pull_requests	40	2
code_security	15	2
context	0	2

Tool-by-Tool Analysis

Tool	Toolset	Tokens	Schema	Rating	Notes
github_context	workflow_context	80	markdown	5	Static, always available
get_teams	context	0	n/a	1	Permission not granted, blocked
get_file_contents	repos	4,200	object	2	README still ~300-entry contributor list before useful content
list_issues	issues	60	array	3	Redacted by integrity policy this run
list_pull_requests	pull_requests	40	array	5	Exactly matched perPage=1 + fields filter
list_workflows	actions	180	object	5	mcpscripts wrapper honored perPage=1 correctly
list_code_scanning_alerts	code_security	15	array	4	Clean empty result
list_discussions	discussions	150	object	4	Cursor pagination, embedded category/author
list_label	labels	2,600	object	3	No perPage control, always full set (~95 of 841)
search_code	search	200	object	4	Succeeded this run, useful match fragments

30-Day Trend Summary

Metric	Value
Data Points	20
Average Daily Tokens	8,700
Average Rating Trend	Stable (3.5 → 3.6)

Recommendations

• High-value tools (rating 4-5): github_context, list_pull_requests, list_workflows (via mcpscripts wrapper), list_code_scanning_alerts, list_discussions, search_code
• Tools needing improvement: get_file_contents (needs excerpt/range support for large files), list_label (needs perPage/filter support — currently always returns the full 841-label set), get_teams (permission friction — consider granting or documenting as unavailable)
• Context-efficient tools (low tokens, high rating): github_context (80 tokens, 5/5), list_pull_requests (40 tokens, 5/5), list_workflows via mcpscripts wrapper (180 tokens, 5/5)
• Context-heavy tools: get_file_contents (4,200 tokens), list_label (2,600 tokens) — both toolsets lack pagination/filtering controls that would let agents request less data

Visualizations

Response Size by Toolset

Usefulness Ratings

Daily Token Trend

Size vs Usefulness

Generated by 🔍 GitHub MCP Structural Analysis · agent · 109.8 AIC · ⌖ 19.5 AIC · ⊞ 11.9K · ◷

• expires on Aug 6, 2026, 4:39 AM UTC-08:00