The Agent Skills Directory

[COMMAND_EXECUTION]: Uses local shell commands such as rg and file existence checks ([ -f ]) to validate project planning files. These commands are restricted to the local file system and are used solely for the stated purpose of planning validation.
[DATA_EXFILTRATION]: Accesses local files (plan.md, tasks.md) to conduct coverage matrices and red-flag scanning. The analysis found no evidence of network operations or attempts to transmit this data to external endpoints.
[PROMPT_INJECTION]: Ingests and processes untrusted user-generated content from specification files. While this creates a surface for indirect prompt injection, the skill's use of specific keyword scanning and interactive remediation helps mitigate the risk.
[SAFE]: Implements a remediation workflow that allows the agent to update local project artifacts. This functionality is consistent with the skill's purpose and is gated by explicit user interaction via the AskUserQuestion mechanism.

review-plan