The Agent Skills Directory

[SAFE]: No security issues were identified in the skill. The skill is entirely descriptive and instructional, consisting of Markdown prompts and YAML configuration files for testing. It does not define or invoke any dangerous tools, scripts, or network operations.
[INDIRECT_PROMPT_INJECTION]: While the skill is designed to process external performance reports, which is a common ingestion surface for indirect prompt injection, it is categorized as safe because it lacks exploitable capabilities.
Ingestion points: The skill accepts external metrics and reports as described in prompts/performance-result-analysis.md and evaluation cases.
Boundary markers: The instructions explicitly command the agent to separate facts from assumptions and state gaps in knowledge, which helps prevent over-reliance on injected content.
Capability inventory: The skill has no file-write, network, or command-execution tools.
Sanitization: The prompt instructs the agent to validate test runs and use masked data or isolated environments for production-related work.

performance-result-analysis