The Agent Skills Directory

[DYNAMIC_CONTEXT_INJECTION]: The skill uses the !command`` syntax in SKILL.md to execute shell commands during the skill loading phase.
Evidence: !d=${CLAUDE_SKILL_DIR}; command -v skill-toc >/dev/null 2>&1 && exec skill-toc "$d"; s=$HOME/.claude/skills/plugin-utils/bin/skill-toc; [ -x "$s" ] && exec "$s" "$d"; echo "(skill-toc unavailable...)" in SKILL.md.
This pattern triggers shell execution without an explicit user action, which is used here to dynamically generate a table of contents but represents a silent execution vector.
[COMMAND_EXECUTION]: The skill instructs the agent to perform various shell commands for archival and cleanup tasks.
Evidence: Instructions recommend using bsdtar for 7z creation and provide a custom Python script for file processing.
[DYNAMIC_EXECUTION]: The skill includes a standalone script that is executed at runtime to modify the content of files provided by the user.
Evidence: scripts/scrub_metadata.py uses the pikepdf library to programmatically edit PDF structures and uses regex to modify OOXML archives.

replication-package