The Agent Skills Directory

[COMMAND_EXECUTION]: The skill invokes the external summarize binary through the Bash tool. It follows security best practices by writing input text to temporary scratch files in the .summarize/ directory instead of direct shell interpolation, which effectively prevents common shell injection vulnerabilities.
[EXTERNAL_DOWNLOADS]: The documentation provides instructions for installing the @steipete/summarize package from the official npm registry or Homebrew. These sources are considered well-known and trusted for package distribution.
[PROMPT_INJECTION]: The skill processes arbitrary text input for summarization, which constitutes an indirect prompt injection surface.
Ingestion points: Untrusted data enters the agent context through stdin, local file paths, or pasted text described in SKILL.md.
Boundary markers: The instructions do not define explicit delimiters to separate the input content from the tool's internal logic.
Capability inventory: The skill uses the Bash tool to execute the summarization CLI.
Sanitization: Shell injection risks are mitigated by the use of file-based input passing rather than variable interpolation into the shell command.

summarize