The Agent Skills Directory

[SAFE]: The skill consists entirely of natural language instructions and structural prompts for orchestrating an AI agent's reasoning process. No malicious code patterns, such as command execution, network exfiltration, or persistence mechanisms, were detected in the analyzed file.- [NO_CODE]: The skill does not include any Python or Node.js scripts, binary executables, or external configuration files. Its logic is implemented via complex system instructions, which reduces the attack surface compared to skills with executable components.- [PROMPT_INJECTION]: The skill demonstrates an indirect prompt injection surface as its primary function involves processing untrusted user data (such as articles, reports, or strategic schemes) to generate inputs for downstream tasks. 1. Ingestion points: Ingests user-provided text as described in the usage scenarios in SKILL.md. 2. Boundary markers: Uses structural Markdown headers and bracketed tags like 【结论】 and 【下游建议】 to delineate output, but lacks specific 'ignore embedded instructions' warnings for external content. 3. Capability inventory: The skill has no access to dangerous capabilities such as file system writes, network requests, or shell execution. 4. Sanitization: No explicit sanitization or filtering logic is defined for the content being processed.

expert-roundtable