The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes data from the Sokosumi API, including job results, task events, and agent input schemas, which represents a potential surface for indirect prompt injection from external agent outputs.
Ingestion points: GET /v1/jobs/:jobId, GET /v1/tasks/:taskId/events, and GET /v1/agents/:agentId/input-schema as documented in SKILL.md.
Boundary markers: None identified in the prompt instructions to delimit external data from agent instructions.
Capability inventory: The skill can hire agents, register coworkers (including setting custom base URLs), and create/monitor tasks using the sokosumi CLI.
Sanitization: No specific sanitization or validation routines for external content are described in the skill instructions.
[EXTERNAL_DOWNLOADS]: The skill instructions include commands to install skill components from the vendor's official GitHub repository.
Evidence: npx skills add https://github.com/masumi-network/sokosumi-cli --skill sokosumi found in SKILL.md.
[COMMAND_EXECUTION]: The skill instructs the agent to execute shell commands using the sokosumi CLI to manage marketplace resources.
Evidence: Multiple command examples in SKILL.md including sokosumi agents list, sokosumi agents hire, and sokosumi coworkers register.

sokosumi