The Agent Skills Directory

[SAFE]: No security issues detected. The skill uses standard markdown instructions and references local documentation and scripts for its functionality. All external resource references trace back to the intended organizational context without signs of typosquatting or malicious intent.- [INDIRECT_PROMPT_INJECTION]: The skill is configured to read company-context.md to provide relevant business guidance. While this creates a potential surface for indirect prompt injection if that file contains untrusted data, this is a standard operational practice for context-aware agents.
Ingestion points: company-context.md (referenced in Context Integration section).
Boundary markers: None explicitly defined in the instructions.
Capability inventory: Internal reasoning using Tree of Thought, role invocation via [INVOKE:role|question], and execution of local strategy scripts.
Sanitization: None explicitly defined for external context reading.

ceo-advisor