The Agent Skills Directory

[SAFE]: The skill implements a legitimate development workflow using the opentaint CLI. All shell commands and file operations are localized to the project environment (specifically within the .opentaint/ directory). The included Python script check-test-result.py is a utility for parsing local JSON results and contains no dangerous network or system operations.
[INDIRECT_PROMPT_INJECTION]: The skill defines a surface for processing untrusted data via the unit and side inputs which are used to construct file paths and shell command arguments. However, this is consistent with the skill's primary purpose as a developer tool for managing project-specific analysis rules.
Ingestion points: User-provided unit and side parameters; tracking files located at .opentaint/tracking/rules/<side>/<unit>.yaml.
Boundary markers: None identified in the provided instructions.
Capability inventory: Execution of shell commands (opentaint, uv run); file read/write operations within the project directory.
Sanitization: None explicitly defined for the unit or side inputs.

create-rule