Saiten MCP Server
Allows the server to fetch hackathon submission data from GitHub issues and post automated feedback comments using the GitHub CLI.
Integrates with GitHub Copilot in VS Code to orchestrate specialized agents for evaluating submissions and generating ranking reports.
Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type
@followed by the MCP server name and your instructions, e.g., "@Saiten MCP Server score all submissions and generate a ranking report"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
Saiten — Agents League @ TechConnect Scoring Agent
Submission Track: 🎨 Creative Apps — GitHub Copilot
License
MIT License — see LICENSE.
Available Tools
5 toolsgenerate_ranking_reportA
Generate a Markdown ranking report and save to reports/ranking.md.
Reads scoring results from data/scores.json and produces a report
containing overall ranking, per-track ranking, and individual
evaluation summaries.
Args:
top_n: Number of top entries to highlight (default: 10).
Returns:
Result dict (report_path, total_scored, top_n, top_entries).
Inputs: top_n. Returns: object
get_scoring_rubricA
Return the scoring rubric for the specified track.
Loads the YAML file ``data/rubrics/{track}.yaml`` and returns
the scoring criteria (name, weight, description, scoring_guide).
Args:
track: Track name. ``"creative-apps"`` | ``"reasoning-agents"``
| ``"enterprise-agents"``
Returns:
Rubric dict with track, track_display_name, criteria (list),
total_weight, score_range, and notes.
Raises:
FileNotFoundError: If the YAML file for the track does not exist.
ValueError: If the track name is invalid.
Inputs: track (required). Returns: object
get_submission_detailA
Fetch detailed submission data for the specified Issue number.
Parses each section of the Issue template and returns scoring data.
GitHub Username is hidden during scoring to eliminate bias, but
retained as the github_username field for report output.
If repo_url points to a GitHub repository, the README is also fetched.
Args:
issue_number: The Issue number to fetch.
Returns:
A dictionary containing detailed submission information.
Raises:
RuntimeError: When gh command execution fails.
Inputs: issue_number (required).
list_submissionsA
Fetch the list of Agents League submissions.
Args:
track: Track name to filter by.
``"creative-apps"`` | ``"reasoning-agents"`` | ``"enterprise-agents"`` | None (all)
state: Issue state. ``"open"`` | ``"closed"`` | ``"all"``
Returns:
A list of submission summaries. Each element is a dictionary
containing issue_number, title, track, project_name, repo_url,
created_at, has_demo.
Raises:
RuntimeError: When gh command execution fails.
Inputs: track, state. Returns: result
save_scoresA
Save scoring results to data/scores.json.
Existing scores for the same Issue are overwritten (idempotent).
New Issues are appended.
Args:
scores: List of scoring result dicts. Each must contain:
- issue_number (int)
- project_name (str)
- track (str)
- criteria_scores (dict[str, int]): per-criterion scores (1-10)
- weighted_total (float): weighted total (0-100)
- evidence (dict[str, str]): per-criterion evidence citations
- confidence (str): 'high', 'medium', or 'low'
- red_flags_detected (list[str]): red flag signals found
- bonus_signals_detected (list[str]): bonus signals found
- strengths (list[str])
- improvements (list[str])
- summary (str)
Returns:
Summary dict (saved_count, updated_count, total_in_store, file_path).
Raises:
OSError: If disk write fails.
Inputs: scores (required). Returns: object
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
TDQS
Scored across 5 tools
Each tool has a clearly distinct purpose with no overlap. generate_ranking_report creates reports, get_scoring_rubric fetches scoring criteria, get_submission_detail retrieves specific submission data, list_submissions lists submissions with filtering, and save_scores stores scoring results. The tools cover different aspects of the scoring workflow without ambiguity.
All tools follow a consistent verb_noun naming pattern (generate_ranking_report, get_scoring_rubric, get_submission_detail, list_submissions, save_scores). The verbs are appropriate and descriptive, and the snake_case convention is maintained throughout without deviation.
With 5 tools, this server is well-scoped for managing a scoring system for submissions. The tools cover essential operations like listing, retrieving details, accessing rubrics, saving scores, and generating reports, which is appropriate for the domain without being excessive or insufficient.
The toolset provides complete coverage for the scoring workflow domain. It includes operations for listing submissions, fetching details and rubrics, saving scores, and generating reports, covering the full lifecycle from data retrieval to output without any obvious gaps that would hinder an agent.
Maintenance
Related MCP Connectors
Run hackathons end to end: events, teams, submissions, judging and winners.
GitHub repo search, stats, language breakdown and comparison as agent tools.
Dependency picker for coding agents: recommends, compares and scores GitHub repos with a verdict.
Manage repositories, users, releases, and automate GitHub workflows
Related MCP Servers
- -licenseBqualityNot gradedmaintenanceEnables AI-driven orchestration of GitHub development workflows including automated issue analysis, code generation, code review, and PR creation through multiple specialized agents. Integrates with GitHub Actions to automate the complete development process from issue to pull request.7-
- AlicenseAqualityBmaintenanceEnables AI coding agents to orchestrate the full software development lifecycle on GitHub, including planning, issue creation, code review, security triage, and release readiness checks.1399 npm1MIT
- AlicenseNot gradedqualityDmaintenanceEnables AI agents to perform GitHub analytics and management tasks, including repository analysis, PR summarization, issue triage, release notes generation, and contributor statistics.39 npmMIT
- FlicenseNot gradedqualityCmaintenanceEnables MCP-compatible LLM clients to query GitHub repository data—commits, issues, contributor stats, and code search—as typed callable tools over stdio.-