The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and analyze untrusted external data such as research papers and scientific claims, which presents a surface for indirect prompt injection attacks. Evidence chain:
Ingestion points: The skill processes research methodology, scientific claims, and study results (typically via the Read tool or user context).
Boundary markers: Absent. There are no instructions to use delimiters or ignore embedded instructions within the scientific texts being analyzed.
Capability inventory: The skill has access to the Bash, Write, and Edit tools as defined in the YAML frontmatter.
Sanitization: Absent. There is no instruction to validate or filter ingested scientific content before it is processed or used to influence agent decisions.
[COMMAND_EXECUTION]: In SKILL.md, the instructions prompt the agent to run a shell command (python scripts/generate_schematic.py) to generate diagrams. However, the referenced script is not included in the 7 files provided, making the actual execution logic and safety of this local dependency unverifiable.

scientific-critical-thinking