budget

budget is a skill for Claude Code from Marcel-Bich/marcel-bich-claude-marketplace. It costs 142 tokens per session (5,324 once invoked), scanned A, original, MIT.

A rulebook for managing two API usage limits: a rolling five-hour limit and a seven-day limit. It also defines a check of commit identity before code is committed.

In plain words
What is it for?
Use it to decide when an agent should continue, slow down, pause, wake up, or stop work based on API usage, and to apply the commit-identity check.
Why use it?
It prevents autonomous work from using the wrong budget or continuing when an API limit is nearly full. It also keeps the current session's context size separate from API usage.

Skill for Claude Code

Written for Claude Code: ${CLAUDE_PLUGIN_ROOT} variable. Also seen: reads .claude/ paths; mentions subagents.

Runs only inside its plugin — its command needs a path that Claude Code sets for a plugin’s own hooks and for nothing else. Install the plugin, not this.

Part of the credo plugin — 26 skills, 15 commands, 6 hooks shipped together

Good fit Use it to decide when an agent should continue, slow down, pause, wake up, or stop work based on API usage, and to apply the commit-identity check.

Compare 6 skills from other repositories ↓
Install

Getting it into your agent

This one installs as part of its plugin. Adding the marketplace and installing the plugin brings it with everything else the plugin ships.

Claude Code
/plugin marketplace add Marcel-Bich/marcel-bich-claude-marketplace
Claude Code
/plugin install credo

Made for: Claude Code.

Or install credo, the plugin that ships this one along with the rest of its 26 skills, 15 commands, 6 hooks.

Wrote 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.

agentmods badge for budget

README.md
[![agentmods](https://agentmods.dev/badge/skills/marcel-bich/marcel-bich-claude-marketplace/budget/github.svg)](https://agentmods.dev/skills/marcel-bich/marcel-bich-claude-marketplace/budget)
Your own site
<a href="https://agentmods.dev/skills/marcel-bich/marcel-bich-claude-marketplace/budget"><img src="https://agentmods.dev/badge/skills/marcel-bich/marcel-bich-claude-marketplace/budget/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.

agentmods 80×15 button for budget

Your own site · 80×15
<a href="https://agentmods.dev/skills/marcel-bich/marcel-bich-claude-marketplace/budget"><img src="https://agentmods.dev/badge/skills/marcel-bich/marcel-bich-claude-marketplace/budget.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 142 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 5,324 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe.
Origin original No closer match found in the catalogue.
Token cost

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.

ModelPer sessionOnce invoked
Fable 5.1 $0.00142 $0.05324
Opus 5 $0.00071 $0.02662
Sonnet 5 $0.00028 $0.01065
Haiku 4.5 $0.00014 $0.00532

Measured 12d ago against content hash c803d5ac634d, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-12, from the pricing page.

Security

Grade A, and why

budget 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 12d 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.

plugins/credo/skills/budget/SKILL.md · 370 lines

How it starts

The opening of the file, as written. The whole thing — 370 lines — stays where its author put it; the contents beside it link to each section on GitHub.

budget - API budget caps, resets, and the commit-identity gate

This skill is the one place that decides how much API budget may be spent and what to do as a limit fills up. It is config-driven: the cap schedule and thresholds live in the credo config, this skill explains how to read and apply them. It also owns the commit-identity gate that must pass before any commit.

Three axes - never confuse them (B11)

There are three completely separate budget axes. Keep them strictly apart; a rule for one never applies to another:

  1. The 5-hour API limit - a rolling window that resets roughly every 5 hours. Field five_hour.utilization (percent used). This skill governs it.
  2. The weekly API limit - a 7-day window. Field seven_day.utilization (percent used). This skill governs it.
  3. The SESSION context fill - how full the current context window is (the compact axis). This is NOT an API budget. It is governed by the compact-plus skill and its compact.thresholds config, not here. Do not mix it into 5h/weekly reasoning.

When you say "budget" be explicit about which axis. "5h at 80 percent" and "context at 80 percent" are unrelated facts with unrelated responses.

Data source: the limit-plugin cache only (B8)

The limit plugin is a PREREQUISITE for all budget data. Without a fresh cache, NO percentage-based cap is measurable or enforceable - not the schedule caps and not the fail-safe caps below, because every one of them is a percentage of a number that only the cache provides. Do not treat the fail-safe caps as a substitute; they too are unenforceable with no fresh cache. The only guardrail enforceable without the cache is a wall-clock timebox.

  • For non-autonomous, ad-hoc reads ("how much budget is left"): budget data is simply unavailable - note that it cannot be read, do not error.
  • For AUTONOMOUS mode entry: never run blind and never silently ignore budgets. The guardrail-availability gate in the credo session-autonomous skill decides what to do - it asks the user (install the limit plugin, run a wall-clock timebox, or proceed at an explicitly accepted risk). See that skill; this skill only supplies the honesty rationale.

Read the full file on GitHub · 370 lines

Changes

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.

  1. 12d ago First seen · 370 lines · 142 tokens per session scan A c803d5ac634d

Subscribe to this mod's changes

budget is a skill published in the GitHub repository Marcel-Bich/marcel-bich-claude-marketplace (13 stars, last pushed yesterday), licensed MIT. It adds 142 tokens to every session and 5,324 once invoked, about $0.0007 per session on Opus 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 2026-08-30.