The Agent Skills Directory

[SAFE]: No malicious patterns, suspicious scripts, or unauthorized network operations were detected. The skill is entirely composed of Markdown-based guidelines for research.
[NO_CODE]: The skill contains no executable code, package manager manifests (like package.json or requirements.txt), or references to external scripts.
[PROMPT_INJECTION]: While the skill involves ingesting data from external web sources which presents a surface for indirect prompt injection, it includes mitigation strategies.
Ingestion points: SKILL.md (instructions to query Google Scholar, USPTO, and government databases).
Boundary markers: Absent for external data streams.
Capability inventory: Reading and synthesizing information from the web.
Sanitization: Mandatory triple-verification (triangulation) of findings and a strict source blacklist to avoid low-quality or untrusted data.
[DATA_EXFILTRATION]: There is no evidence of unauthorized data access or exfiltration. The skill references well-known and official domains (e.g., stats.gov.cn, flk.npc.gov.cn) exclusively for information retrieval.

interdisciplinary-research