The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes the 'gh' (GitHub CLI) and 'curl' tools to retrieve information from the official WordPress development environment.
[INDIRECT_PROMPT_INJECTION]: The skill possesses a surface for indirect prompt injection as it processes data from external sources. The skill implements a selective ingestion strategy by specifically targeting comments from 'github-actions[bot]' to build its base props list. Mandatory evidence chain: Ingestion points include pull request comments (GitHub API) and ticket content from core.trac.wordpress.org; Boundary markers are absent for the final interpolated message; Capability inventory includes read/write access to GitHub PRs and local shell execution; Sanitization is not explicitly defined for the extracted metadata.
[REMOTE_CODE_EXECUTION]: An automated alert was triggered by a command piping remote content to an interpreter. Evidence: 'curl -s "https://core.trac.wordpress.org/ticket/{TICKET_NUMBER}" | python3 -c "..."'. Analysis: This is determined to be a false positive as the Python script is a local string defined within the skill, and the remote URL points to a well-known project resource providing HTML data for parsing rather than executable code.

core-commit