The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted user input through the $ARGUMENTS variable to determine which maintenance workflow to execute.
Ingestion points: $ARGUMENTS variable parsed within the <process> block in SKILL.md.
Boundary markers: No explicit delimiters or instructions to ignore instructions embedded within arguments are present in the top-level SKILL.md.
Capability inventory: The skill is configured to use Bash, Write, and Edit tools to perform system updates and modify local configuration files.
Sanitization: The skill parses the first token for specific flags (--sync, --reapply) but passes the remaining unvalidated tokens directly to sub-workflows.
[EXTERNAL_DOWNLOADS]: The skill fetches update packages and changelog information from the npm registry. This behavior is standard for the tool's stated objective of keeping software up to date.
[COMMAND_EXECUTION]: The skill uses the Bash tool to run internal workflows and manage files within the ~/.claude/gsd-core/ directory, facilitating local synchronization and patching operations.

gsd-update