The Agent Skills Directory

[SAFE]: The skill serves as a reference and instructional guide for researchers. It consists of theoretical econometrics content, workflow branches (DID, IV, RDD, Selection, Structural), and references to standard statistical tools.
[PROMPT_INJECTION]: No evidence of prompt injection, role-play bypasses, or instructions to ignore safety filters was detected. The instructions are aligned with academic and professional research standards.
[DATA_EXPOSURE_&_EXFILTRATION]: No hardcoded credentials, sensitive file path access, or unauthorized network operations were identified. The content is strictly methodological.
[OBFUSCATION]: The file was scanned for Base64, zero-width characters, homoglyphs, and hidden text patterns. No obfuscation techniques were found.
[UNVERIFIABLE_DEPENDENCIES_&_REMOTE_CODE_EXECUTION]: The skill does not perform external package installations or execute remote code. It references local documentation and standard statistical estimators (e.g., Callaway-Sant'Anna, Sun-Abraham) which are expected in the context of financial research tools.
[INDIRECT_PROMPT_INJECTION]: The skill describes a workflow that ingests user-provided research designs for auditing. While this creates an indirect injection surface, there are no instructions that would cause the agent to deviate from safe execution.
Ingestion points: User-provided manuscript details and econometric specifications.
Boundary markers: None explicitly defined in the Markdown text.
Capability inventory: The skill references tool calls for statistical testing (e.g., detect_design, rdrobust, mccrary_test), but these are restricted to the agent's pre-defined statistical environment.
Sanitization: Not applicable as the skill is primarily informational.
[DYNAMIC_CONTEXT_INJECTION]: No use of shell-execution placeholders (!command) was found in the SKILL.md file.

jfe-identification