The Agent Skills Directory

[COMMAND_EXECUTION] (HIGH): The skill explicitly directs the agent to use sed -i and awk for in-place file modifications. These commands are populated with data from external sources such as 'security findings' or 'code review' results. This pattern facilitates command injection or malicious code insertion if an attacker can influence the findings used by the agent.
[DATA_EXFILTRATION] (LOW): The skill performs broad codebase analysis using git diff and grep, and reads various project documentation files for context, granting the agent extensive read access to the repository.
[PROMPT_INJECTION] (LOW): The skill exhibits an indirect prompt injection surface (Category 8). (1) Ingestion points: README.md, CONTRIBUTING.md, and outputs from other tools. (2) Boundary markers: Absent. (3) Capability inventory: sed and awk file modification. (4) Sanitization: The instructions do not specify any sanitization of external content before it is interpolated into shell commands.

create-todos