The Agent Skills Directory

[COMMAND_EXECUTION]: The file assets/example_code_vulnerable.py contains a demonstration function command_injection_example that uses subprocess.run(command, shell=True). This is explicitly presented as a vulnerable code pattern for educational purposes and is not part of the skill's operational scripts.
[REMOTE_CODE_EXECUTION]: The file assets/example_code_vulnerable.py includes an insecure_deserialization_example function using the pickle.loads() method. It also defines a MaliciousClass designed to execute rm -rf / via os.system upon deserialization. These components are documented as demonstration material for security training.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and analyze external project code, creating a surface for indirect prompt injection if an attacker embeds malicious instructions within the source files being reviewed.
Ingestion points: Project source files identified in step 1 of the review process (SKILL.md).
Boundary markers: Absent; the report is generated as standard Markdown without specific isolation markers for code snippets.
Capability inventory: The skill uses Read, Write, Edit, Grep, Glob, and Bash tools to analyze files and generate reports.
Sanitization: No specific sanitization or escaping steps are defined for handling content extracted from the reviewed codebase before including it in the output report.

performing-security-code-review