[mcp-analysis] MCP Structural Analysis - 2026-09-11

[github-actions[bot]]

Tested one representative tool from each of 9 GitHub MCP toolsets plus the injected workflow-context block against github/gh-aw. Today's total: 7,275 tokens across 10 calls, average usefulness 3.60/5. Best: github_context and list_workflows (both 5/5). Worst: list_issues (2/5, blocked by an integrity-policy redaction on every measured day so far). The code_security and repos toolsets remain the most expensive on average over the 30-day window; pull_requests and issues remain the cheapest.

Full Structural Analysis Report

Executive Summary

Metric	Value
Tools Analyzed	10
Total Tokens (Today)	7,275
Average Usefulness Rating	3.60/5
Best Rated Tool	github_context / list_workflows: 5/5
Worst Rated Tool	list_issues: 2/5

Usefulness Ratings for Agentic Work

Tool	Toolset	Rating	Assessment
github_context	workflow_context	⭐⭐⭐⭐⭐	Excellent — zero-cost injected identity block
list_workflows (mcpscripts)	actions	⭐⭐⭐⭐⭐	Excellent — clean JSON, explicit pagination, no bloat
list_code_scanning_alerts	code_security	⭐⭐⭐⭐	Good — actionable remediation markdown, but verbose
list_pull_requests	pull_requests	⭐⭐⭐⭐	Good — flat, minimal, directly actionable
search_code	search	⭐⭐⭐⭐	Good today — inline code fragment avoided a follow-up call
get_file_contents	repos	⭐⭐⭐	Adequate — icon/_meta overhead roughly doubles payload
list_discussions	discussions	⭐⭐⭐	Adequate — good pagination metadata, but _meta-dominated
get_label	labels	⭐⭐⭐	Adequate — tiny payload buried in fixed _meta overhead
search_users	users	⭐⭐⭐	Adequate — minimal identity fields only
list_issues	issues	⭐⭐	Limited — redacted item blocks assessment (13/13 days)

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	3	content.resource.text, _meta.serverInfo	Base64 icon block ~half the payload
list_issues	array	2	number, title, state, labels	One item redacted by integrity policy
list_pull_requests	array	1	number, title, state, html_url	No _meta overhead
list_workflows	object	2	workflows[].id/name/path/state, total_count	mcpscripts wrapper, clean schema
list_code_scanning_alerts	array	3	rule.id/severity/full_description/help, location	Schema inconsistent across scanners (CodeQL vs 3rd-party)
list_discussions	object	3	discussions[].category/user, pageInfo, totalCount	Good relationships, _meta-heavy
get_label	object	1	name, color, description, id	Tiny payload, _meta-heavy
search_users	object	2	items[].login/id/profile_url	No name/bio/role fields
search_code	object	2	items[].path, text_matches[].fragment	Inline snippet reduces follow-up calls

Response Size Analysis (30-Day Average by Toolset)

Toolset	Avg Tokens	Tools Tested
code_security	2,294	13
repos	2,096	13
labels	903	13
discussions	688	13
search	629	13
users	615	13
actions	171	13
workflow_context	93	13
issues	62	13
pull_requests	38	13

Tool-by-Tool Analysis (Today)

Tool	Toolset	Tokens	Schema	Rating	Notes
github_context	workflow_context	90	markdown	⭐⭐⭐⭐⭐	Zero-cost identity context
get_file_contents	repos	1,660	object	⭐⭐⭐	.gitignore unchanged since 2026-09-04
list_issues	issues	65	array	⭐⭐	1 issue redacted, 0 visible
list_pull_requests	pull_requests	48	array	⭐⭐⭐⭐	Single actionable PR record
list_workflows	actions	172	object	⭐⭐⭐⭐⭐	mcpscripts wrapper, best-in-class
list_code_scanning_alerts	code_security	2,600	array	⭐⭐⭐⭐	5 open CodeQL alerts, all with remediation examples
list_discussions	discussions	700	object	⭐⭐⭐	totalCount 10,609
get_label	labels	610	object	⭐⭐⭐	"bug" label
search_users	users	630	object	⭐⭐⭐	1 user matched
search_code	search	700	object	⭐⭐⭐⭐	Snippet included inline, no rate-limit today

30-Day Trend Summary

Metric	Value
Data Points	130
Unique Days Measured	13
Average Daily Tokens	7,589
Average Rating Trend	Stable (3.3–3.7 range; dips on 09-02/09-08/09-09 tied to search_code rate-limit errors, not schema regressions)

Recommendations

• High-value tools (rating 4–5): github_context, list_workflows (mcpscripts wrapper), list_pull_requests, list_code_scanning_alerts, search_code (when not rate-limited).
• Tools needing improvement: list_issues — the integrity-policy redaction has now hidden one item on all 13 measured days, meaning an agent can never fully assess repo issues via this path; search_code remains prone to secondary rate-limit errors (3 of 13 days), which zeroes its usefulness intermittently.
• Context-efficient tools (low tokens, high rating): github_context (90 tokens, 5/5), list_workflows (172 tokens, 5/5), list_pull_requests (48 tokens, 4/5).
• Context-heavy tools: code_security and repos toolsets remain the two most expensive on average (2,294 and 2,096 tokens/call over 30 days) — largely driven by verbose CodeQL remediation markdown and the fixed ~2.2k-char base64 icon block repeated in nearly every non-wrapper response.

Visualizations

Response Size by Toolset

Usefulness Ratings

Daily Token Trend

Size vs Usefulness

Warning

Firewall blocked 1 domain

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

• api.anthropic.com

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

network:
  allowed:
    - defaults
    - "api.anthropic.com"

See Network Configuration for more information.

Generated by 🔍 GitHub MCP Structural Analysis · claude · agent · 157.4 AIC · ⌖ 10.1 AIC · ⊞ 11.5K · ◷

• expires on Sep 12, 2026, 4:02 AM UTC-08:00