The Agent Skills Directory

[SAFE]: The skill is a standard implementation of the Harness CCM service integration. It defines instructions for the agent to interact with the vendor's official MCP server using approved tools like harness_get and harness_list.
[INDIRECT_PROMPT_INJECTION]: The skill ingests data from external cloud provider billing and recommendation systems via the Harness platform, which represents a potential surface for indirect prompt injection. This is a low-risk architectural observation given the trusted nature of the data source and the restricted scope of the tool capabilities.
Ingestion points: Cloud cost summaries, breakdowns, and recommendations retrieved via harness_get and harness_list tools in SKILL.md.
Boundary markers: The skill does not implement specific delimiters or instructions to ignore instructions embedded within the retrieved billing data.
Capability inventory: The skill is primarily focused on data retrieval, with the additional capability to report feedback on anomalies using the harness_execute tool.
Sanitization: No explicit sanitization or validation of external tool outputs is requested within the instructions.

analyze-costs