The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied business hypotheses and experiment results to generate an 'Experiment Card'. Ingestion points: User responses during the Build-Measure-Learn cycle in SKILL.md. Boundary markers: None. Capability inventory: None; the skill only generates textual advice and cards. Sanitization: None.
[METADATA_POISONING]: The skill includes external links and metadata referring to the author (deciqai) and reputable sources like HBR and OpenAI. No deceptive or hidden instructions were found in the metadata fields.
[SAFE]: No obfuscation, unauthorized dependencies, or persistence mechanisms were detected. The skill's behavior matches its stated purpose.

lean-startup