The Agent Skills Directory

[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it extracts and executes verification commands from untrusted repository files.
Ingestion points: The skill reads AGENTS.md, README files, package scripts, and other repository guidance to identify verification commands.
Boundary markers: Absent. There are no instructions to use delimiters or ignore instructions embedded within the data files.
Capability inventory: The skill is capable of executing arbitrary shell commands discovered during the repository scan.
Sanitization: Absent. The skill does not validate or sanitize the commands found in the repository before execution.
[COMMAND_EXECUTION]: The skill performs shell command execution as part of its core functionality.
Evidence: It uses git fetch, git merge, and git branch commands.
Evidence: It executes 'documented verification commands' found in the local environment, which can include arbitrary scripts or binaries defined in the codebase.

update-branch