claude-code-academic-workflow

Warn

Audited by Gen Agent Trust Hub on Jun 17, 2026

Risk Level: MEDIUMEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill instructs users to clone a repository from an unverified third-party GitHub account (pedrohcgs/claude-code-my-workflow). This introduces external scripts that are not from a verified vendor.
  • [COMMAND_EXECUTION]: The workflow relies on executing local shell scripts provided in the external repository, specifically ./scripts/validate-setup.sh and ./scripts/save-context-snapshot.sh. It also executes various CLI tools including xelatex, Rscript, quarto, git, and gh. The documentation also suggests the use of sudo fc-cache -fv for font management, which involves elevated system privileges.
  • [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it ingests and processes untrusted research manuscripts, R scripts, and Quarto documents while possessing execution capabilities. Ingestion points: Processes user-provided .tex, .R, .qmd, and .yaml files. Boundary markers: No explicit delimiters or instructions are provided to prevent the model from following malicious commands that might be embedded within the documents being analyzed. Capability inventory: Access to the shell, document compilers, statistical software (R), and version control systems. Sanitization: There is no evidence of sanitization or content validation for the external files processed by the skill.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Jun 17, 2026, 06:52 PM
Security Audit — agent-trust-hub — claude-code-academic-workflow