The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to execute various shell commands to run, test, and regenerate tools, specifically using the python interpreter for modules like tooluniverse.cli, tooluniverse.generate_tools, and custom scripts like scripts/test_new_tools.py.
[COMMAND_EXECUTION]: The skill involves a dynamic execution pattern where the agent modifies Python source files in the src/tooluniverse/ directory and subsequently executes them using pytest or tool-specific wrappers.
[PROMPT_INJECTION]: An Indirect Prompt Injection surface exists because the agent is directed to ingest and act upon data from untrusted external sources, such as online API documentation and agent-reported bug descriptions.
Ingestion points: External official API documentation retrieved via search, and bug reports/feedback provided by other agents.
Boundary markers: No specific boundary markers or instructions to disregard embedded commands in the documentation are provided in the skill instructions.
Capability inventory: The skill has the capability to write to the local filesystem (modifying src/ and tests/ directories) and execute those modified files via shell commands.
Sanitization: There are no documented procedures for validating, sanitizing, or escaping content found in the external documentation before it is translated into code changes.
[SAFE]: The skill promotes secure credential management by instructing that API keys should be accessed via environment variables (os.environ.get()) rather than being hardcoded or passed as tool parameters.

devtu-fix-tool