The Agent Skills Directory

[SAFE]: The skill enforces a mandatory human-in-the-loop workflow by requiring explicit user confirmation for all outreach dispatches. It specifically instructs the agent to draft content first and only send once the user provides clear approval.
[SAFE]: Credential management for the REVOR_API_KEY is handled through standard local environment files located at ~/.config/RevorSkill/.env. The skill includes specific safeguards to prevent the leaking of these secrets, such as masking keys in output and prohibiting storage in logs or memory files.
[INDIRECT_PROMPT_INJECTION]: The skill has a potential attack surface because it processes external recipient data and user-provided context to generate messages.
Ingestion points: The agent ingests recipient contact details (profile URLs, email addresses) and user-defined objectives from the chat context (SKILL.md).
Boundary markers: The instructions explicitly demand a 'draft-only' phase and forbid the fabrication of contact data (SKILL.md).
Capability inventory: The skill uses MCP tools and REST API calls to revor.ai to dispatch messages and like LinkedIn posts (reference/dispatch/dispatch.md).
Sanitization: The skill validates the can_send status of accounts and verifies that channel-specific required fields are present before execution (reference/accounts.md).
[SAFE]: The skill performs legitimate network operations to the vendor's domain (revor.ai) to check account readiness and execute dispatch jobs as intended by the primary functionality.

revor-outreach