The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it reads and processes untrusted data from GitHub issue descriptions and comments. This data directly influences the implementation plan generated by the planning team. \n
Ingestion points: SKILL.md (Step 1) and references/workflow.md (Section 2) use gh issue view to retrieve the issue body and all comments. \n
Boundary markers: The workflow does not specify clear delimiters or instructions to the agent to ignore potentially malicious commands embedded in the issue text. \n
Capability inventory: The skill has extensive capabilities, including executing shell commands (git, gh), running arbitrary repository test suites (npm test, pytest, etc.), and pushing code to remote repositories. \n
Sanitization: There is no evidence of sanitization or filtering of the content retrieved from GitHub. \n- [COMMAND_EXECUTION]: The skill relies on executing various shell commands to manage the development workflow. This includes git for branch and commit management, gh for GitHub interactions, and various build/test tools discovered within the target repository. \n- [EXTERNAL_DOWNLOADS]: The skill references and facilitates the installation of the team-executor skill via npx skills add thatjuan/agent-skills --skill team-executor. While this is a vendor-owned resource from the skill author 'thatjuan', it involves downloading and installing external code into the agent's environment.

implement-issue