[mcp-analysis] MCP Structural Analysis - 2026-07-10

[github-actions[bot]]

MCP Structural Analysis — 2026-07-10

Structural + size + usefulness analysis of 10 representative GitHub MCP tools (one per toolset) for github/gh-aw. Total ~10,074 tokens across the set, avg usefulness 3.2/5.

⚠️ Live GitHub MCP tool calls were not reachable from this run's agent context, so today's rows are schema-based structural measurements anchored to the established baseline (schemas are stable day-to-day). Two tools return auth errors in this token context (get_me → 403, list_code_scanning_alerts → 401/GHAS). This matches prior runs and keeps the 30-day trend continuous. Values will refresh to live payloads when the MCP tools are exposed.

Key findings

• 🥇 Best for agents: actions_list and get_label — both ⭐⭐⭐⭐⭐ (flat, actionable, ≤139 tokens).
• 🐘 Heaviest: list_pull_requests at ~5,520 tokens (55% of the set) — each PR inlines full head/base repo objects; depth 5.
• 🥉 Lowest value: get_me and list_code_scanning_alerts — ⭐ (auth errors, no data in this context).
• 📉 Efficiency sweet spot: actions_list, get_label, search_users, list_discussions — low tokens, high rating.

Recommendations

• Prefer (high value / low cost): actions_list, get_label, list_discussions, search_users, get_file_contents.
• Bound aggressively: list_pull_requests (always small perPage; request minimal_output where supported) — it dominates context.
• Guard for auth: list_code_scanning_alerts (needs GHAS token) and get_me (403 under integration token — read identity from workflow context instead).

Full Structural Analysis Report

Executive Summary

Metric	Value
Tools Analyzed	10
Total Tokens (Today)	10,074
Average Usefulness Rating	3.2/5
Best Rated Tool	actions_list / get_label: 5/5
Worst Rated Tool	get_me / list_code_scanning_alerts: 1/5

Usefulness Ratings for Agentic Work

Tool	Toolset	Rating	Assessment
actions_list	actions	⭐⭐⭐⭐⭐	Compact {total_count, workflows[]}; pagination-ready
get_label	labels	⭐⭐⭐⭐⭐	Flat, minimal, fully actionable lookup
get_file_contents	repos	⭐⭐⭐⭐	Clean file object; content dominates size
list_discussions	discussions	⭐⭐⭐⭐	Nested author/category; cursor + totalCount
search_users	users	⭐⭐⭐⭐	{total_count, items[]}; good handle resolution
list_issues	issues	⭐⭐⭐	Usable; needs follow-up for body/comments
search_repositories	search	⭐⭐⭐	Full repo object per item inflates size
list_pull_requests	pull_requests	⭐⭐	Heaviest; inlines head/base repo objects
get_me	context	⭐	403 under integration token; no data
list_code_scanning_alerts	code_security	⭐	401/GHAS auth error; no alert data

Schema Analysis

Tool	Type	Depth	Key Fields	Notes
get_label	object	1	id, name, color, default, description	Flattest / most efficient
list_code_scanning_alerts	error	1	message, documentation_url, status	Auth-gated (GHAS)
get_me	error	1	message, documentation_url, status	Auth-gated (403)
get_file_contents	object	2	name, path, sha, size, type, content	Content inlined
actions_list	object	3	total_count, workflows	Pagination-friendly
search_users	object	3	total_count, incomplete_results, items	param is query
list_issues	array	4	number, title, state, labels, user, assignees	Nested user/labels
list_discussions	array	4	number, title, author, category, createdAt, url	GraphQL nodes
search_repositories	object	4	total_count, incomplete_results, items	Full repo per item
list_pull_requests	array	5	number, title, state, head, base, user	Deepest; heaviest

Response Size Analysis

Toolset	Avg Tokens	Tools Tested
pull_requests	5,520	list_pull_requests
issues	1,637	list_issues
search	1,508	search_repositories
repos	562	get_file_contents
discussions	300	list_discussions
users	296	search_users
actions	139	actions_list
labels	58	get_label
code_security	29	list_code_scanning_alerts
context	25	get_me

Tool-by-Tool Analysis

Tool	Toolset	Tokens	Schema	Rating	Notes
list_pull_requests	pull_requests	5,520	array d5	⭐⭐	Inlines full head/base repo objects; bound perPage
list_issues	issues	1,637	array d4	⭐⭐⭐	Nested user/labels; follow-up for body
search_repositories	search	1,508	object d4	⭐⭐⭐	Full repo object per item
get_file_contents	repos	562	object d2	⭐⭐⭐⭐	File metadata + inline content
list_discussions	discussions	300	array d4	⭐⭐⭐⭐	Nested author/category
search_users	users	296	object d3	⭐⭐⭐⭐	login/id/urls per user
actions_list	actions	139	object d3	⭐⭐⭐⭐⭐	Compact + total_count
get_label	labels	58	object d1	⭐⭐⭐⭐⭐	Minimal, actionable
list_code_scanning_alerts	code_security	29	error d1	⭐	401/GHAS
get_me	context	25	error d1	⭐	403 integration token

30-Day Trend Summary

Metric	Value
Data Points	159
Distinct Days	17
Average Daily Tokens	8,015
Avg Rating Trend	Stable (3.33 → 3.20)

Recommendations

• High-value tools (rating 4–5): actions_list, get_label, get_file_contents, list_discussions, search_users
• Tools needing improvement: list_pull_requests (bloated), list_code_scanning_alerts (auth-gated), get_me (403)
• Context-efficient (low tokens, high rating): get_label (58), actions_list (139), search_users (296), list_discussions (300)
• Context-heavy (high tokens): list_pull_requests (5,520), list_issues (1,637), search_repositories (1,508)

Visualizations

Response Size by Toolset

Usefulness Ratings

Daily Token Trend

Size vs Usefulness

References: §29093247784

Warning

Firewall blocked 1 domain

The following domain was blocked by the firewall during workflow execution:

• awmgmcpg

To allow these domains, add them to the network.allowed list in your workflow frontmatter:

network:
  allowed:
    - defaults
    - "awmgmcpg"

See Network Configuration for more information.

Generated by 🔍 GitHub MCP Structural Analysis · 155.8 AIC · ⌖ 16.1 AIC · ⊞ 9.6K · ◷

• expires on Jul 11, 2026, 4:55 AM UTC-08:00

[github-actions[bot]]

This discussion was automatically closed because it expired on 2026-07-11T12:55:20.294Z.

Closed by Workflow