The Agent Skills Directory

[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection (Category 8) by design. It fetches issue descriptions and comments from GitHub, which are untrusted external data sources, and specifically instructs the agent to analyze and 'adopt' suggestions from these comments.\n
Ingestion points: GitHub issue content and comments fetched via gh issue view and gh api commands in SKILL.md.\n
Boundary markers: Absent. The instructions do not provide delimiters or warnings to ignore malicious instructions embedded within the issue data.\n
Capability inventory: The agent has extensive shell capabilities including gh CLI for repository interaction, bd for task management, and pnpm for building/testing code. It also has the ability to delegate tasks to specialized subagents.\n
Sanitization: Absent. There is no evidence of filtering, validation, or escaping of the content fetched from GitHub before it is processed or used to influence code changes.\n- [COMMAND_EXECUTION]: The skill relies heavily on shell command execution to perform its primary functions. It uses the GitHub CLI (gh) for remote repository management and a local task manager (bd). It also executes build-time commands such as pnpm type-check and pnpm build. While these are part of the intended workflow, executing build commands on a codebase that may have been modified based on untrusted input from issue comments increases the risk profile.

process-issues