Caps fallback tier to prevent expensive AI model escalation during retries.
npx skills add https://github.com/AesopScott/mojo --skill retry-cap
Please help me install this Agent Skill. Skill: retry-cap Source: https://github.com/AesopScott/mojo/tree/main/harnesses/skills/retry-cap Command: npx skills add https://github.com/AesopScott/mojo --skill retry-cap
This Skill prevents silent, expensive escalations in AI model usage by capping the fallback tier during retry operations.
Use the retry-cap skill to inspect the current fallback configuration and apply a cap to prevent escalation to pricier models.
High-intent search queries and answers about installing and using this skill.
To prevent expensive model fallback during AI retry operations, apply a strict cap on fallback tier escalation. This enforces defined boundary rules to ensure retries do not automatically switch to pricier models, keeping token expenditure within budget.
Fallback tier escalation in model management is the process of switching to alternative AI models when a primary model fails. Without a cap, this mechanism can silently escalate to significantly more expensive models, causing unexpected token cost spikes.
Enforce cost efficiency boundaries for prompt engineering workflows by inspecting your fallback configuration and applying a cap. This restricts model selection during retry cycles, verifying cost-saving outcomes through positive and boundary testing.
Yes, you can inspect and modify existing retry fallback logic using a structured runbook. This harness inspection process allows you to review current configurations and apply strict boundaries to prevent silent cost escalations.
Token usage spikes when a primary AI model fails because the system automatically escalates to a higher, more expensive fallback tier. Applying a cap restricts this escalation, ensuring retries stay within defined cost limits.
No specific dependencies are required to cap fallback tiers. The system operates independently to inspect your model management configuration and enforce strict boundary rules for cost efficiency during retry operations.
xenitV1Switch AI agent behavior between brainstorm, implement, debug, review, teach, ship, and orchestrate modes.
AesopScottEnforce token usage ceilings per route in AI dispatch harnesses.
latestaiagentsAnalyze LLM API token usage and provide cost optimization strategies.
planetaryescapeRoute user messages to optimal AI models via task classification and cost-aware tier selection.
