The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill identifies a surface for processing external cloud data that could theoretically contain untrusted strings.
Ingestion points: The skill directs the agent to analyze "cloud bills", "Cost and Usage Reports (CUR)", and "billing exports" which include user-controlled metadata like resource tags and names as mentioned in SKILL.md and references/cost_visibility_allocation.md.
Boundary markers: The instructions do not define specific delimiters for separating billing data from agent instructions, relying on the agent's native handling of these data types.
Capability inventory: The skill focuses on analysis and reporting workflows using the agent's core capabilities, without adding custom tools or scripts.
Sanitization: Validation of resource metadata is treated as a platform-level concern, as the skill provides high-level analysis frameworks.
[NO_CODE]: The skill consists entirely of instructional markdown files and does not package any executable scripts, binaries, or configuration files that execute code.

finops-analyst