The Agent Skills Directory

[PROMPT_INJECTION]: The skill exhibits a vulnerability surface for indirect prompt injection through its automated file generation logic.
Ingestion points: The script scripts/generate_deploy.py accepts user-supplied strings via command-line arguments (--name, --env, --owner) and interpolates them directly into the output file deploy-plan.md.
Boundary markers: There are no delimiters or explicit instructions in the generated file to differentiate between data and instructions, making it possible for an attacker to inject headers or commands that the agent might follow.
Capability inventory: The skill is granted significant permissions, including Bash, Write, Edit, and Read tools, which could be exploited if the agent obeys instructions injected into the generated plan.
Sanitization: No sanitization or validation is performed on the input arguments before they are written to the file system.

deployment-engineer