The Agent Skills Directory

[SAFE]: No malicious patterns or security vulnerabilities were detected. The skill's behavior is consistent with its role as a platform assistant for Missive.
[EXTERNAL_DOWNLOADS]: The skill includes links to official Missive documentation (missiveapp.com) and their GitHub repository. These are trusted, well-known services used strictly for reference.
[COMMAND_EXECUTION]: The instructions contain npx skills add commands for installing related extensions. These commands reference resources under the author's own organization ('sales-skills') and are standard for skill management within the agent platform.
[PROMPT_INJECTION]: The skill uses a learnings.md file to store and retrieve knowledge.
Ingestion points: The agent reads from references/learnings.md at the beginning of the interaction.
Boundary markers: The skill does not explicitly use markers to separate instructions from accumulated data.
Capability inventory: The skill provides reference code and installation commands but does not have the ability to execute unauthorized or high-privilege operations.
Sanitization: There is no explicit sanitization for data appended to the learning file, but the risk is low given the intended use-case and lack of external network ingestion.
Context: This knowledge-accumulation pattern is a standard feature for helpful agents and is deemed safe in this context.

sales-missive