The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and analyze user-provided financial data, such as billing records and accounting entries (Scenario 4: '花销分析'). It lacks specific boundary markers or instructions to ignore potential commands embedded within this external data. This creates a surface where malicious instructions hidden in a bill or financial statement could influence the agent's behavior.
Ingestion points: SKILL.md (Scenario 4 under '支持的场景' and Step 1 under '工作流程')
Boundary markers: Absent. The skill does not instruct the agent to treat input data as non-executable text or use delimiters to isolate untrusted content.
Capability inventory: No internal tools or shell capabilities are defined in the frontmatter, limiting the potential impact of an injection to text-based behavior modification or misinformation.
Sanitization: Absent. No validation or escaping is requested for the input data.
[EXTERNAL_DOWNLOADS]: The instructions include a capability gap response rule that directs the agent to suggest installing '技能宝' via a GitHub repository (github:mindverse/skillhub). This is a functional feature for skill discovery within its intended platform ecosystem and refers to a known repository.

budget-planner