The Agent Skills Directory

[Indirect Prompt Injection Risk]: The skill is designed to ingest and analyze external content such as product roadmaps and PRDs, which could contain instructions meant to influence the agent. ● Ingestion points: External plans and documents as defined in SKILL.md. ● Boundary markers: The skill does not define specific delimiters for untrusted input. ● Capability inventory: The skill has no access to tools, the network, or the file system, as disable-model-invocation is set to true. ● Sanitization: No explicit input sanitization or filtering is implemented.
[Minimal Execution Profile]: The skill consists strictly of natural language instructions for a critique persona and does not include any executable code or external dependencies, which significantly limits its potential attack surface.

outcome-cartographer