The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses the handover CLI and MCP interface to execute various state management commands, including handover whoami, handover search, handover show, handover pull, and handover continue (SKILL.md).
[EXTERNAL_DOWNLOADS]: The skill retrieves project artifacts and revisions from the Handover vendor's service using the handover pull command (SKILL.md).
[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted content from external handover records which could be manipulated to contain malicious instructions.
Ingestion points: External data is retrieved and processed through handover.search, handover show, and handover pull (SKILL.md).
Boundary markers: No explicit delimiters are specified to separate the external data from the agent's instructions.
Capability inventory: The skill can write files to a local directory via handover pull and publish new project revisions via handover continue (SKILL.md).
Sanitization: The instructions do not require the agent to sanitize or validate the content of retrieved artifacts or revision notes before processing.

handover-resume