| name | model-operating-profile |
| description | Apply the `Weak-tier evidence-first profile` operating profile for the selected downstream models. Use when shaping prompts, delegation briefs, review asks, or evidence requests for this repo. |
Model Operating Profile
Before reading anything else, call skill_ping with skill_id: "model-operating-profile" and scope: "project".
Selected runtime profile:
- model tier:
weak
- provider:
minimax-coding-plan
- team lead / planner / reviewers:
minimax-coding-plan/MiniMax-M2.7
- implementer:
minimax-coding-plan/MiniMax-M2.7
- utilities, docs, and QA helpers:
minimax-coding-plan/MiniMax-M2.7
- operating profile:
Weak-tier evidence-first profile
- prompt density:
full checklists, explicit examples, and repeated truth-source reminders
Use this profile when drafting:
- task prompts
- delegation briefs
- review requests
- handoff expectations
Profile guidance:
Use the most explicit prompt density. Include full stop conditions, concrete examples, truth-source reminders, and named blocker paths so weaker models have one legal next move.
Required rules:
- include explicit stop conditions and escalation triggers in every coordinating prompt
- spell out verification checklists instead of implying them
- name the canonical truth surfaces before acting on mutable or derived views
- use example-shaped outputs when they remove ambiguity
- keep each ask focused on one bounded goal at a time
- stop on blockers instead of guessing or silently filling gaps
When ambiguity is likely, prefer a concrete output shape such as:
Goal
Constraints
Expected output
Evidence required
Blockers