The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes a local Python script scripts/parallel_cost.py to calculate API costs. The script runs offline and operates on local JSON files or standard input.
[INDIRECT_PROMPT_INJECTION]: The instructions in SKILL.md warn the agent to treat search results, excerpts, and task outputs as untrusted content, preventing the execution of instructions embedded in retrieved web data.
[DYNAMIC_EXECUTION]: The scripts/test_helper.py file uses importlib to dynamically load the parallel_cost.py module for testing purposes. This is limited to local files within the skill directory.
[CREDENTIALS_UNSAFE]: The skill provides explicit instructions to the agent to never print the PARALLEL_API_KEY, mitigating the risk of accidental credential exposure in logs or user interactions.

parallel-efficiency