The Agent Skills Directory

[SAFE]: No obfuscation, hardcoded credentials, or unauthorized network operations were detected in the skill files.
[SAFE]: The skill does not execute external code, download remote dependencies, or request elevated privileges.
[SAFE]: The execution logic is limited to text analysis within a subagent context, matching the stated purpose of identifying themes in research notes.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to process untrusted data in the form of paper notes, which presents a theoretical attack surface for indirect prompt injection.
Ingestion points: The paper_notes input variable defined in SKILL.md and processed in prompt.md.
Boundary markers: Absent; the input is interpolated into the prompt without explicit delimiters or instructions to ignore embedded commands.
Capability inventory: None. The skill does not authorize any tools or execute shell commands, significantly mitigating the risk of injection-led harm.
Sanitization: No input validation or sanitization is specified for the paper_notes content.

thematic-coding