Trigger structured consultation when implementation obstacles require tradeoff alignment.
npx skills add https://github.com/davidlee/spec-driver --skill consult-davidlee
Please help me install this Agent Skill. Skill: consult Source: https://github.com/davidlee/spec-driver/tree/main/.spec-driver/skills/consult Command: npx skills add https://github.com/davidlee/spec-driver --skill consult-davidlee
It prevents silent drift when unanticipated obstacles, tradeoffs, or emergent complexity would cause impactful architectural or agent-driven decisions to be made without alignment.
If you hit an unexpected blocker, tell the AI to stop, summarize the current context and discoveries, list the options and tradeoffs, and ask for guidance before continuing.
High-intent search queries and answers about installing and using this skill.
To prevent architectural drift, you must stop and trigger structured consultation when unanticipated obstacles emerge. This forces a pause to summarize findings, present tradeoffs, and request guidance before proceeding with risky adaptations.
Scope control during implementation tradeoffs requires surfacing multiple options and their impacts before committing. This ensures decisions that alter planned functionality are explicitly aligned rather than silently diverting the scope.
Decision alignment for agent governance is needed when an agent proposes adding new coupling, substantial debugging, or policy ambiguity. It acts as a guardrail to prevent unsafe, agent-driven decisions without explicit contextual summarization.
Risk escalation for unsafe architectural changes requires halting the task, summarizing the current plan and discoveries, listing options with tradeoffs, and asking a guidance-seeking question before proceeding.
No, when research or debugging threatens to divert scope, structured consultation is required. It enforces a stop condition to contextualize the obstacle and align on the tradeoff before allowing the implementation to adapt.
Stop conditions for decision alignment only trigger consultation; they do not resolve the obstacle. You must still manually summarize context, present explicit tradeoffs, and formulate a guidance-seeking question to proceed safely.
quanganh208Analyze technical and architectural questions for system design reviews.
special-place-administratorPlan and execute multi-file refactorings with automatic rollback on verification failure.
ZpankzRoute complex tasks to appropriate subagent types using trigger conditions and complexity scoring.
khapn-2933Analyze proposed changes for architectural, security, performance, and UX risks.
