The Agent Skills Directory

[COMMAND_EXECUTION]: Provides shell commands to generate a local script at ~/.claude/briefing.sh and set its execution permissions using chmod +x.
[COMMAND_EXECUTION]: Instructs the user to modify their crontab to achieve persistence for the daily briefing script.
[COMMAND_EXECUTION]: References the use of lunchy-go to install a persistent launchd service on macOS.
[PROMPT_INJECTION]: The skill presents an indirect prompt injection surface as it processes untrusted data from multiple integrated skills.
Ingestion points: Data enters the agent context from c-email, c-slack, c-github, and other skills.
Boundary markers: The briefing template does not utilize specific delimiters or instructions to ignore embedded commands within the ingested content.
Capability inventory: The generated briefing is used as input for the claude CLI tool via a scheduled script.
Sanitization: No evidence of filtering, escaping, or validation of the external content is present.

c-briefing