The Agent Skills Directory

[COMMAND_EXECUTION]: The skill explicitly recommends using GIT_SSL_NO_VERIFY=1 when running in sandbox environments. This environment variable disables TLS certificate verification for Git-based operations (including the gh CLI), which facilitates Man-in-the-Middle (MitM) attacks, allowing an attacker to intercept or modify communication with GitHub.
[COMMAND_EXECUTION]: The skill fetches external content from GitHub Project draft items (titles and bodies) and interpolates them directly into shell commands: gh issue create --title "<ドラフトのタイトル>" --body "<ドラフトの本文>". If the draft content contains shell-sensitive characters (like backticks, semicolons, or command substitutions) and is not properly handled by the agent, it could lead to arbitrary command execution on the host system.
[COMMAND_EXECUTION]: The skill processes untrusted external data with the following risk factors (Indirect Prompt Injection surface):
Ingestion points: Data is ingested via gh project item-list in SKILL.md at Step 1 and Step 4.
Boundary markers: The instructions do not define boundary markers or instruct the agent to ignore instructions embedded within the draft titles or bodies.
Capability inventory: The skill possesses write capabilities through gh issue create, gh project item-add, and gh api (POST).
Sanitization: There are no instructions or scripts provided to sanitize the external content before it is used to construct API calls or shell commands.

project-create-issues