[mcp-analysis] MCP Structural Analysis - 2026-08-28

[github-actions[bot]]

Daily structural analysis of GitHub MCP tool responses across 9 toolsets plus the injected workflow-context block. 10 data points collected today (2026-08-28), 40 total across the 4-day rolling window. Average usefulness held steady at 3.6/5. Best: github_context and list_workflows (mcpscripts wrapper), both 5/5. Worst: list_issues (2/5) — a fourth consecutive day where one issue is silently redacted by an integrity policy filter with no visible content, which continues to block agents from assessing it. The clearest recurring pattern: several non-wrapper tools (get_label, list_discussions, search_users, search_code, get_file_contents) pay a fixed ~550-token base64 icon _meta.serverInfo overhead on every call, which can exceed the size of the actual payload for small lookups.

Full Structural Analysis Report

Executive Summary

Metric	Value
Tools Analyzed	10
Total Tokens (Today)	4,262
Average Usefulness Rating	3.6/5
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 pagination metadata, no bloat
list_code_scanning_alerts	code_security	⭐⭐⭐⭐	Good, trivially small and unambiguous
list_pull_requests	pull_requests	⭐⭐⭐⭐	Good, flat and directly actionable
search_code	search	⭐⭐⭐⭐	Good — text_matches fragment now includes an actual snippet
get_file_contents	repos	⭐⭐⭐	Adequate — icon overhead now roughly equals content size
get_label	labels	⭐⭐⭐	Adequate — overhead dominates a ~30-token payload
list_discussions	discussions	⭐⭐⭐	Adequate — good relationships, overhead-heavy
search_users	users	⭐⭐⭐	Adequate — minimal identity fields only
list_issues	issues	⭐⭐	Limited — redacted item blocks assessment (4th day running)

Schema Analysis

Tool	Type	Depth	Key Fields	Notes
github_context	markdown	1	actor, repository, workspace, workflow-run-id	Flat bullet list, no MCP call
list_workflows	object	2	total_count, per_page, workflows[].id/name/path/state	Explicit pagination metadata
list_code_scanning_alerts	array	1	(empty today)	Rich schema (rule, severity, location) when populated
list_pull_requests	array	1	number, title, state, html_url	Flat single-item array
search_code	object	2	items[].name/path/text_matches[].fragment	Fragment includes real code snippet
get_file_contents	object	3	content.resource.text, _meta.serverInfo	Deepest nesting, icon _meta included
get_label	object	1	name, color, description, id	Shallow but icon-overhead heavy
list_discussions	object	3	discussions[].category.name/user.login, pageInfo, totalCount	Nested category/user relations
search_users	object	2	items[].login/id/profile_url	No name/bio, needs follow-up call
list_issues	array	2	number, title, state, labels	One item redacted by integrity filter

Response Size Analysis

Toolset	Avg Tokens (4-day)	Tools Tested
repos	2,230	4
labels	970	4
discussions	670	4
search	592	4
users	564	4
code_security	533	4
actions	181	4
workflow_context	103	4
issues	60	4
pull_requests	36	4

Tool-by-Tool Analysis (Today, 2026-08-28)

Tool	Toolset	Tokens	Schema	Rating	Notes
github_context	workflow_context	90	markdown, depth 1	5	Zero MCP cost, immediately actionable
list_workflows	actions	172	object, depth 2	5	mcpscripts wrapper, clean pagination, no icon bloat
list_code_scanning_alerts	code_security	10	array, depth 1	4	No open critical alerts today
list_pull_requests	pull_requests	35	array, depth 1	4	Single open PR, flat and minimal
search_code	search	720	object, depth 2	4	Snippet fragment reduces follow-up calls
get_file_contents	repos	1,250	object, depth 3	3	Icon overhead ≈ content size for a small file
get_label	labels	610	object, depth 1	3	~30-token payload, ~550-token fixed overhead
list_discussions	discussions	680	object, depth 3	3	Same fixed icon overhead pattern
search_users	users	630	object, depth 2	3	Minimal identity fields, needs follow-up
list_issues	issues	65	array, depth 2	2	Integrity-filtered issue hides content (4th day)

30-Day Trend Summary

Metric	Value
Data Points	40 (4 days collected so far)
Average Daily Tokens	5,938
Average Rating Trend	Stable (3.6 → 3.7 → 3.6 → 3.6)

Recommendations

• High-value tools (rating 4-5): github_context (free), list_workflows via the mcpscripts wrapper, list_pull_requests, list_code_scanning_alerts, search_code — prefer these when the schema fits the task.
• Tools needing improvement: list_issues — the integrity-policy redaction has now recurred for 4 consecutive days with zero visible content, and should either surface a summarized/sanitized view or a clearer signal of what triggered the filter.
• Context-efficient tools (low tokens, high rating): github_context, list_workflows, list_pull_requests, list_code_scanning_alerts.
• Context-heavy tools: get_file_contents (repos, 2,230 avg tokens) is consistently the heaviest — prefer directory listings with fields filtering over full-file fetches, and note the fixed ~550-token icon _meta block inflates every non-wrapper call (get_label, list_discussions, search_users, search_code) regardless of payload size. A wrapper that strips _meta.serverInfo.icons (following the list_workflows/list_label mcpscripts pattern) would meaningfully cut overhead across the board.

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 · 111.8 AIC · ⌖ 8.77 AIC · ⊞ 11.5K · ◷

• expires on Aug 29, 2026, 8:04 AM UTC-08:00

[github-actions[bot]]

This discussion was automatically closed because it expired on 2026-08-29T16:04:41.409Z.

Closed by Workflow