The Agent Skills Directory

[SAFE]: The skill provides procedural instructions for research workflows and does not include any executable code, sensitive file access, or network exfiltration commands.
[PROMPT_INJECTION]: Indirect prompt injection surface identified: 1. Ingestion points: The skill ingests untrusted data from external sources via web-search, web-research, and literature-engine SOPs. 2. Boundary markers: No specific delimiters or safety instructions are provided to the agent regarding the handling of embedded instructions in the ingested data. 3. Capability inventory: This skill file itself defines no dangerous capabilities like file writing or subprocess execution. 4. Sanitization: No sanitization or input validation of external content is specified.

scoping-survey