The Agent Skills Directory

[COMMAND_EXECUTION]: The script scripts/mutate.py executes a user-specified command (such as a test runner) via subprocess.run() to evaluate the effect of code mutations. This is a core component of the mutation testing workflow.
[DYNAMIC_EXECUTION]: The mutation process involves programmatically altering local source code and executing it to verify the robustness of verification gates. The scripts/mutate.py script identifies mutation sites (operators and numbers), applies changes, and verifies the outcome by running a secondary process.
[INDIRECT_PROMPT_INJECTION]: The tool ingests source code from the local file system, creating an attack surface for indirect prompt injection if a targeted file contains malicious content. 1. Ingestion points: scripts/mutate.py reads files from the --target path. 2. Boundary markers: The results are reported in plain text without explicit delimiters. 3. Capability inventory: The script can read/write files and execute subprocesses. 4. Sanitization: The script uses the tokenize library to restrict mutations to specific code tokens (operators and numbers), preventing modification of string literals or comments.

gating