github-stars-mcp
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
| GITHUB_PERSONAL_ACCESS_TOKEN | No | GitHub personal access token. If not provided, the GitHub CLI (`gh`) must be authenticated with `user` scope. |
Instructions
Guidance the server publishes about itself, which clients place ahead of the tool catalog so the model reads it before choosing anything.
This server publishes no instructions, or was last inspected before Glama recorded them.
Capabilities
Features and capabilities supported by this server
Protocol revision2025-11-25
| Capability | Details |
|---|---|
| tools | {} |
Tools
Functions exposed to the LLM to take actions
| Name | Description |
|---|---|
| github_orchestrate_workersA | Map-Reduce Step 1: Takes an immutable frozen snapshot of all starred repos (eliminating mid-run mutation drift), extracts baseline taxonomy, and generates balanced worker chunk plans. |
| github_get_worker_chunkA | Worker Tool: Fetches the exact frozen snapshot chunk assigned to this worker by session_id and worker_id, including pre-distilled READMEs. |
| github_submit_worker_digestA | Map-Reduce Step 2: Allows sub-agents to submit their distilled category & pitch digest. Validates against baseline taxonomy, prevents reassigned worker race conditions, and handles GraphQL list sync fallback. |
| github_force_reduce_sessionA | Resilience Tool: Force-reduces an orchestration session when a worker is orphaned, times out, or fails. Can proceed with available repos or reassign missing chunks. |
| github_get_orchestration_statusB | Inspect current progress, completed workers, worker state machine, and compiled categories of an orchestration session. |
| github_get_user_infoA | Get authenticated GitHub user details and total count of starred repositories (cached with 15min TTL). |
| github_list_starredB | Fetch a paginated list of starred repositories with details (cached with 10min TTL). |
| github_get_readmeB | Fetch the README markdown for a repository with semantic distillation, ETag caching (24h TTL), and context budget controls. |
| github_batch_get_starred_with_readmeB | High-performance batch tool: fetches starred repositories + distilled READMEs with dynamic token budgeting, local caching, and speculative next-batch prefetching. |
| github_clear_cacheB | Clear memory and disk caches for GitHub stars and READMEs. |
| github_get_user_listsB | Get all native GitHub Star Lists created by the authenticated user along with their current items and IDs. |
| github_create_user_listA | Create a new native GitHub Star List on github.com (requires token with 'user' scope). |
| github_assign_repo_to_listsA | Assign or update a repository's membership in native GitHub Star Lists (requires token with 'user' scope). |
| github_export_catalogB | Save or update an organized Markdown catalog file (e.g. GITHUB_STARS.md) grouping repositories by category with summaries and tags. |
| github_audit_stars_healthA | Audit health, maintenance freshness, and licensing safety across starred repositories. Detects dead abandonware, stale projects, and unmaintained dependencies with composite 0-100 health scoring. |
| github_search_starsB | Instant offline concept & problem-to-solution discovery across starred repositories. Powered by zero-dependency local hybrid BM25 and weighted field scoring (< 5ms). |
| github_recommend_stackB | AI Agent Tech-Stack Recommender: recommends vetted starred tools and libraries for a project brief or task description, with copy-paste installation commands and rationale. |
Prompts
Interactive templates invoked by user choice
| Name | Description |
|---|---|
No prompts | |
Resources
Contextual data attached and managed by the client
| Name | Description |
|---|---|
No resources | |
TDQS
Scored across 17 tools
Most tools have clear, distinct responsibilities, especially the orchestration pipeline steps. However, list_starred and batch_get_starred_with_readme overlap significantly in retrieving starred repos, and get_readme also overlaps with the batch tool's README fetching, creating potential confusion.
All tools follow a consistent github_verb_noun pattern in snake_case. Even longer names like batch_get_starred_with_readme and assign_repo_to_lists fit the convention, making the set highly predictable.
At 17 tools, the set is slightly above the ideal 3-15 range but each tool serves a defined purpose—the orchestration group alone accounts for 5. The count feels heavy but not bloated for the breadth of features offered.
The server covers star retrieval, README reading, list management, catalog export, health auditing, and search. Notable gaps include no way to remove a star/unstar, no delete or rename for user lists, and list membership only being add/update without explicit removal functionality.