The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes Bash to execute gh (GitHub CLI) and git commands. It follows security best practices by properly quoting variables (e.g., "$REPO") and employing heredocs with quoted delimiters (<<'EOF') when constructing issue bodies. This approach effectively prevents accidental command substitution or shell injection from the content of the issues.
[INDIRECT_PROMPT_INJECTION]: The skill interacts with untrusted external data by reading issue titles and descriptions from GitHub during triage and duplicate detection operations. While this constitutes a surface for indirect prompt injection, the risk is low as the data is used for classification and display rather than being executed as instructions or code.
[DATA_EXPOSURE]: The skill accesses the local .git/config to determine the repository origin (git remote get-url origin). This is a legitimate operation for the skill's stated purpose and does not involve exfiltration of sensitive credentials.

github-issues