The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill possesses a surface for indirect prompt injection by design, as its primary function involves ingesting untrusted data from various sources to generate output.
Ingestion points: The skill reads data from external sources including test logs (github/get_job_logs), file contents for TODOs/docs (github/get_file_contents), and user reports or audits (github/issue_read).
Boundary markers: There are no explicit boundary markers or instructions to the agent to disregard hidden commands within the ingested content, although it does use a structured quality checklist.
Capability inventory: The agent is empowered to create new repository issues (github/create_issue) and add comments to existing ones (github/add_issue_comment).
Sanitization: The skill includes a manual "Safety" instruction in its checklist to avoid copying secrets, credentials, or personal data. However, there is no mention of programmatic sanitization or escaping of external content to prevent instructions in those sources from being executed by the agent.
[COMMAND_EXECUTION]: The skill utilizes a set of GitHub-specific tools to perform its duties, including github/search_issues, github/search_pull_requests, github/issue_read, github/get_file_contents, github/search_code, github/actions_list, github/get_job_logs, github/list_label, github/create_issue, github/update_issue_labels, and github/add_issue_comment. These tools are standard for repository management and are provided by the vendor.

keep-the-pool-full