The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill interacts exclusively with api.github.com, which is a well-known and trusted service for repository management.
[DATA_HANDLING]: The read_file function in scripts/github_client.py ingests content from external GitHub repositories. While this content is added to the agent's context, the skill does not implement specific boundary markers or sanitization, creating a potential surface for indirect prompt injection if a user targets a malicious repository.
Ingestion points: read_file function in scripts/github_client.py.
Boundary markers: Absent; content is returned as a raw string.
Capability inventory: File writing and pull request creation via commit_file, commit_files, and create_pull_request in scripts/github_client.py.
Sanitization: Absent; content is not escaped or validated before processing.
[COMMAND_EXECUTION]: The skill implementation relies on the standard urllib library for network operations and does not use subprocess, os.system, or other shell execution vectors.
[CREDENTIALS_UNSAFE]: No hardcoded credentials were found. The skill provides clear instructions for users to manage their GitHub Personal Access Tokens securely using the platform's Project Knowledge feature.

invoking-github