The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external content (URLs, page copy) to perform rewrites and has filesystem write permissions. This creates a surface where malicious instructions hidden in the source content could influence the agent's behavior.\n
Ingestion points: SKILL.md specifies using draft copy or external URLs as source material for the optimization process.\n
Boundary markers: Absent. Instructions do not specify the use of delimiters or provide warnings for the agent to ignore instructions potentially embedded in the source material.\n
Capability inventory: The skill is granted Read and Write tool permissions in SKILL.md to analyze and modify files.\n
Sanitization: Absent. No evidence of content validation or filtering is provided in the skill instructions.

content-optimizer