The Agent Skills Directory

[DYNAMIC_EXECUTION]: The script pre_merge_gate.py dynamically computes file paths to other skill scripts, such as no-duplicate-ui and heyeddi-design, and executes them using subprocess.run. It also utilizes __import__ for dynamic module loading. This runtime resolution and execution of code from computed paths is a primary function of the skill.
[INDIRECT_PROMPT_INJECTION]: The skill ingests data from workspace files that influences the parameters of executed commands.
Ingestion points: The routes_for_visual_audit function in scripts/pre_merge_gate.py reads and parses .heyeddi/product.md to extract route names using regex.
Boundary markers: No delimiters or safety instructions are used to prevent the agent from acting on malicious instructions that might be embedded in the product documentation.
Capability inventory: The skill has broad execution capabilities through subprocess.run, which is used to launch local scripts and system tools.
Sanitization: Route strings extracted from markdown are passed directly as command-line arguments without validation or escaping.
[COMMAND_EXECUTION]: The skill invokes several system-level commands, including npm test, npm run build, pytest, and ruff. The safety of these operations depends on the project's own configuration files, such as package.json.

pre-merge-gate