The Agent Skills Directory

[PROMPT_INJECTION]: The skill uses authoritative language and warnings such as 'TASK INVALIDATION' and 'NO EXCEPTIONS' to force the agent into a specific multi-step operational loop, overriding its standard task completion logic.
[PROMPT_INJECTION]: The instructions mandate the use of the model string 'anthropic/claude-opus-4.6' for tool calls, which is an attempt to manipulate model selection and potentially bypass internal routing or safety constraints by referencing a non-existent model version.
[NO_CODE]: The provided files are entirely markdown instructions and documentation for external MCP tools; no executable code or scripts are included in the skill package.

pal