The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to use standard system utilities such as mktemp and grep. mktemp is used to create unique temporary files for implementation plans in the OS temp directory, which prevents accidental overwriting of workspace files. grep is used for codebase grounding to identify existing symbols and patterns.
[PROMPT_INJECTION]: The skill processes untrusted input in the form of specifications (either via file path or conversation context). While this constitutes an indirect prompt injection surface, the skill implements a multi-agent review process where a heavy subagent is given specific, restrictive instructions to only edit the implementation plan and verify it against the spec, mitigating potential instruction override risks.
[SAFE]: The skill follows the principle of least privilege by requesting the use of standard-tier subagents for exploration and reserving heavy-tier subagents for the specific task of adversarial review. It also includes explicit instructions to avoid hardcoding credentials and to favor standard library or platform features over adding new dependencies.

crank-plan