GitHub operations via PyGithub (no subprocess) - FastMCP-based GitHub operations for Claude Code
Critical gaps across all 18 tools. No input schemas visible in provided source code, cannot verify parameter types, constraints, or descriptions. Tool descriptions are present but generic (10-50 chars), lacking WHEN/WHY context. No output schemas documented. No error handling guidance. Tools like 'github_auto_commit_and_pr' and 'github_full_merge_cycle' violate single-responsibility principle (combine multiple operations). Naming is verb-first (good), but descriptions do not explain dependencies, prerequisites, or recovery paths. No evidence of parameter validation, enum constraints, or field-level documentation.
Add a comment to a GitHub issue or pull request
Automatically commit changes and create a pull request
Close an existing GitHub issue
Create a new GitHub issue
Create a branch for a GitHub issue
Create a new label in a GitHub repository
Create a new milestone in a GitHub repository
No input schemas visible in source code. Cannot verify parameter types, constraints, enums, or descriptions. All 18 tools scored 0 for schema.
Tool descriptions are generic and under 50 characters. Missing WHEN/WHY context, prerequisites, and recovery guidance. LLMs cannot determine when to select each tool vs alternatives.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 37 | 2026-07-28+ | v2 |
Create a new GitHub pull request
Execute a full merge cycle for a GitHub pull request
Get details of a specific GitHub issue
Get the status of a GitHub pull request
Add labels to a GitHub issue
List comments on a GitHub issue or pull request
List GitHub issues
Merge a GitHub pull request
Reopen a closed GitHub issue
Update an existing GitHub issue
Validate a GitHub build or CI status
Two tools combine multiple responsibilities: 'github_auto_commit_and_pr' (commit + PR creation) and 'github_full_merge_cycle' (validate + merge + cleanup). Violates single-responsibility principle. Should split into separate tools.
No output schemas documented. LLMs cannot plan downstream tool calls or extract required fields (e.g., issue_id, pr_number) for chaining.
No error handling guidance visible. No recovery hints, retryability classification, or actionable error messages. Agents cannot self-correct on failures.