The Agent Skills Directory

[SAFE]: The skill's primary function is to analyze and challenge software designs or plans provided by the user. While it has access to shell tools (Bash) and file system tools (Read, Grep, Glob), their usage is explicitly scoped by the instructions to 'exploring the codebase' to verify facts rather than asking the user. This is a standard and safe practice for AI-assisted development tools.- [PROMPT_INJECTION]: The skill exhibits an attack surface for indirect prompt injection because it processes untrusted user-supplied plans.
Ingestion points: Input is received via $ARGUMENTS and by reading plan files or descriptions as specified in SKILL.md.
Boundary markers: There are no explicit delimiters (e.g., XML tags or triple backticks) or instructions to the agent to ignore embedded commands within the processed data.
Capability inventory: The skill is configured with access to Read, Grep, Glob, Bash, and Agent tools.
Sanitization: The skill does not implement validation or escaping for the input data. However, given the intended use case of plan analysis, this is a known surface with low risk of exploitation.

grill-me