work-within-usage-limitslisted

Plan and checkpoint long builds so a usage/context limit never destroys work in flight. Use when the user mentions a usage limit, token limit, "finish it before the limit hits", asks you to monitor remaining budget, asks to split a project into portions sized to the remaining budget, or says "Continue from where you left off". Also use unprompted at the start of any build expected to run more than an hour, or any multi-agent workflow — those are the two things that get cut off mid-flight and lose everything. Covers portioning, durable checkpoints, resumable state, and recovering findings from agents that died before reporting.
hellokianben-collab/vishal-agarwal-context · ★ 0 · AI & Automation · score 60

Install: claude install-skill hellokianben-collab/vishal-agarwal-context

# Work within usage limits The user's most-repeated operational constraint, stated directly: > *"I want to create the final version of it… my main problem is your usage limit. I want you to > monitor how much usage limit do you have left? and each time divide the project of creating the > final version in portions that you can complete using the remaining usage limit. **By doing this we > will not waste any usage.**"* And separately: *"just finish it quickly before the usage limit hits"*, *"token limit is 30k"*. The failure mode is not running out. It is running out **with the work only in the transcript**. --- ## Rule 1 — a portion is a thing that survives the session ending Do not split by "phases of the code". Split by **what would still be true if the session died right now**. | Bad portion | Good portion | |---|---| | "Write the backend" | "One endpoint, deployed, returning 200 on the live URL" | | "Design the schema" | "Migration written and run; `\d table` shows it" | | "Investigate the bug" | "Findings written into the skill's gotcha section" | Every portion ends with something **on disk or on a URL**. Never with something only in context. ## Rule 2 — checkpoint before you think you need to Order of operations at the end of every portion: 1. Land the artifact (deploy / write the file / run the migration). 2. **Write the state down** — update the project skill or a `STATE.md`. 3. Only then start the next portion. If the budget is visibly tightening, invert