The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to execute an external script named check-no-excuse-rules.py to identify modularity violations. This script is not part of the skill's own files, meaning the agent will execute whatever file exists with that name in the target repository. If an attacker places a malicious script in the codebase, it will be executed with the agent's privileges.- [PROMPT_INJECTION]: The skill exhibits an indirect prompt injection attack surface by processing untrusted source code and passing it to subagents for cleanup.
Ingestion points: The skill reads source code from the repository identified via git diff or provided as explicit arguments.
Boundary markers: Absent; the skill passes the file content directly into subagent prompts without delimiters or warnings to ignore embedded instructions.
Capability inventory: The skill has the capability to modify files (Edit tool), execute shell commands (git, awk, and the project's test suite), and spawn sub-tasks using the Task tool.
Sanitization: No sanitization or content validation is performed on the source code before it is processed by the agent.

remove-ai-slops