The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to use operating system standard commands (such as open on macOS or start on Windows) to automatically display the generated HTML report in the user's browser after completion.
[EXTERNAL_DOWNLOADS]: The HTML template used for reports fetches typography resources from Google Fonts' official domains (fonts.googleapis.com and fonts.gstatic.com).
[INDIRECT_PROMPT_INJECTION]: The skill possesses a surface for indirect prompt injection because it ingests and transforms untrusted project notes which could potentially contain malicious instructions meant to override the agent's summary logic.
Ingestion points: User-provided raw notes and work records (SKILL.md, README.md).
Boundary markers: None explicitly defined to separate untrusted data from the transformation prompt.
Capability inventory: The agent has the capability to write files to the local desktop and execute shell commands to open them.
Sanitization: The skill relies on the agent's NLP logic to extract specific business fields (What/Why/Result) rather than performing programmatic sanitization.

boss-view