The Agent Skills Directory

[DYNAMIC_EXECUTION]: The skill provides shell command templates in references/agent-gather-calendar.md and references/agent-gather-github-attendee-context.md that interpolate user-supplied variables like TARGET_DATE. This design presents a risk of command injection if the agent does not strictly validate or sanitize these inputs before execution.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from multiple external sources to generate meeting briefs. 1. Ingestion points: The skill fetches web search results for candidate research (agent-gather-web-candidate-context.md), Slack thread snippets, and Fireflies meeting summaries. 2. Boundary markers: The instructions lack explicit delimiters or instructions to the agent to treat fetched content as data rather than instructions, potentially allowing embedded malicious content to influence agent behavior. 3. Capability inventory: The skill is authorized to write files to the local disk, which could be abused if an injection succeeds. 4. Sanitization: There are no requirements for sanitizing or escaping the external content before it is rendered into the final markdown templates.
[PERSISTENCE]: The skill contains a dedicated resource, references/cron-recipes.md, which provides detailed instructions for users to configure recurring automated runs using macOS launchd and Claude Code's scheduling features, enabling persistent execution on the host machine.

p-meeting-prep