The Agent Skills Directory

[PROMPT_INJECTION]: The skill facilitates multi-agent communication by allowing subagents to read and write to a shared 'blackboard' filesystem. This creates a surface for Indirect Prompt Injection, as an agent processing untrusted external data could write instructions into the blackboard that are subsequently executed or obeyed by other agents in the chain.
Ingestion points: SKILL.md describes the use of cc_blackboard_read and the cc://blackboard/ resource to load findings into an agent's context.
Boundary markers: Absent. The provided prompt templates in SKILL.md do not include delimiters or warnings to ignore instructions contained within the findings.
Capability inventory: SKILL.md explicitly allows the Bash and Read tools, providing high-privilege execution capabilities if an injection is successful.
Sanitization: Absent. There is no evidence of sanitization or validation of the content within the findings field before it is consumed by downstream agents.

orchestration-blackboard