The Agent Skills Directory

[PROMPT_INJECTION]: The skill has an inherent attack surface for indirect prompt injection because it processes untrusted data from user bug reports and codebase contents.\n
Ingestion points: User-provided bug descriptions (Step 1) and codebase content/git history (Step 2) enter the agent's context.\n
Boundary markers: The skill does not define specific delimiters to separate untrusted data from agent instructions.\n
Capability inventory: The skill uses the gh issue create command to post data to GitHub and an exploration subagent to read files.\n
Sanitization: The instructions explicitly forbid the inclusion of file paths and implementation details in the GitHub issue, which helps prevent the exfiltration of sensitive internal code structure through the tool output.\n- [COMMAND_EXECUTION]: The skill uses git log and gh issue create to perform its primary functions. These are standard tools for development workflows and are used here within their intended scope.

triage-issue