The Agent Skills Directory

[SAFE]: The skill was evaluated using an assume-malicious posture and no risky patterns, obfuscation, or malicious instructions were found. It serves exclusivel y as a design reference.
[NO_CODE]: This skill consists entirel y of markdown guidelines and does not include or download any program matic scrip t s, binaries, or package dependencies.
[INDIREC T_PRO M P T_I N J E C T I O N]: The skill ingests untruste d data in the form of interface copy to be refined. Ingestion point s: The '[target]' argumen t in SKIL L.md. Boundary marker s: Absen t; the skill does not use delimiter s to separate use r input from instruction s. Capabilit y inventor y: None; the skill does not posses s any privilege d tool s, network capabilities, or file system access. Sanitizatio n: Absen t; no escaping or validation is perform ed on the input text, thoug h this pose s minima l risk due to the absenc e of exploitable tool s.

clarify