Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add Team-Deepiri/deepiri-axiom --skill stay-within-limitsgit clone --depth 1 https://github.com/Team-Deepiri/deepiri-axiomWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/team-deepiri/deepiri-axiom/stay-within-limits)<a href="https://agentmods.dev/skills/team-deepiri/deepiri-axiom/stay-within-limits"><img src="https://agentmods.dev/badge/skills/team-deepiri/deepiri-axiom/stay-within-limits/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/team-deepiri/deepiri-axiom/stay-within-limits"><img src="https://agentmods.dev/badge/skills/team-deepiri/deepiri-axiom/stay-within-limits.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00041 | $0.00575 |
| Opus 5.5 | $0.00016 | $0.00230 |
| Sonnet 5.5 | $0.00008 | $0.00115 |
| Haiku 4.5 | $0.00004 | $0.00057 |
Grade A, and why
stay-within-limits scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 11d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 68 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Stay within limits
Keep long-running work inside the host's usage windows (e.g. 5-hour and weekly). Check before substantial work and between parallel waves. At ≥95% of an active window, pause new launches until safe to continue.
Source: BuilderIO/skills (stay-within-limits). MIT.
Core loop
- Bounded wave — default ≤3 parallel subagents unless told otherwise
- Wait for the wave to finish (don't kill in-flight work just to save budget)
- Check current usage with the host's budget/usage tool
- If ≥95% on any active window — stop new work; schedule a self-contained resume
- On resume, re-check the real window (do not trust wall-clock alone)
Usage signals
Prefer a first-party host usage tool when available.
Claude Code example:
npx -y ccusage@latest blocks --active --json
Use active block start, cost/percentage, and time remaining. A new block-start timestamp beats "enough time passed."
If only cost is reported, map through the known account limit. User-configured earlier caution thresholds are fine; default stop remains 95%.
Cursor / other hosts: use whatever usage/billing UI or CLI the environment provides; same pause/resume rules apply.
Pausing and resuming
When a wake/resume tool exists, schedule:
min(3600, secondsUntilWindowClears)
Chain wakeups for longer waits. Each wakeup: re-check usage → reschedule if still hot → continue only when safely below threshold.
Wake prompts must be self-contained: remaining plan, 95% rule, wave throttle, usage command, previous window id, next verification, next handoff packets.
Choosing wait mechanism
- Wake/resume tool — instructions attached to future resume
- Background sleep/watcher — fixed timers / observable process events
- Cron — only for recurring fresh-session work
Avoid short-interval polling for events the host already notifies. Prompt-cache miss after a long sleep is acceptable; preserving the limit matters more.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 11d ago First seen · 68 lines · 41 tokens per session scan A 66c197a3cec4
stay-within-limits is a skill published in the GitHub repository Team-Deepiri/deepiri-axiom (2 stars, last pushed yesterday), licensed Apache-2.0. It adds 41 tokens to every session and 575 once invoked, about $0.0002 per session on Opus 5.5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-10-26.
stay-within-limits compared with skills from other repositories
| Mod | Type | Stars | Last push | Licence | Scan | Tokens | Works with |
|---|---|---|---|---|---|---|---|
| stay-within-limitsTeam-Deepiri/deepiri-axiom | Skill | 2 | yesterday | Apache-2.0 | A | 41 | Claude Code |
| orbit-notionnexu-io/open-design | Skill | 106k | yesterday | Apache-2.0 | A | 117 | Claude Code, Codex |
| agentmailpaperclipai/paperclip | Skill | 111k | yesterday | MIT | A | 30 | Claude Code, Codex |
| Cortexdanielmiessler/LifeOS | Skill | 20k | yesterday | MIT | A | 184 | Claude Code |
| rowboat-spacesrowboatlabs/rowboat | Skill | 18k | yesterday | Apache-2.0 | A | 60 | Claude Code, Codex |
| feishuHmbown/CodeWhale | Skill | 41k | 6d ago | MIT | A | 33 | Claude Code, Codex |
| pinchtab-mcppinchtab/pinchtab | Skill | 11k | 10d ago | MIT | A | 52 | Claude Code, Codex |