Skip to content

Performance: GitHub API rate limit exhaustion and severe bottleneck during repository analysis #337

Description

@Shreesh-Sree

Description

When analyzing candidates with a large number of GitHub repositories, the agent makes deep API calls to fetch commit history, languages, and metadata for every single repository. This creates two critical issues:

  1. Severe performance bottlenecks (evaluations take an unacceptably long time).
  2. High risk of hitting GitHub API rate limits, which causes the agent to crash or fail to complete the evaluation.

Expected Behavior

The agent should prioritize the most relevant repositories to analyze and cap the maximum number of repositories to prevent rate limit exhaustion and ensure timely processing.

Steps to Reproduce

  1. Run python score.py on a candidate's resume/profile who has 50+ repositories on GitHub.
  2. Observe the console hanging for a long period during the GitHub retrieval phase.
  3. If run multiple times or with multiple high-repo candidates, observe GitHub API Rate Limit errors.

Proposed Solution

Modify the repository retrieval logic to sort the candidate's repositories by stargazers_count (descending) and cap the deep analysis to the top 15 repositories. This ensures the agent still evaluates the most impactful work while preserving API quota and execution speed.

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions