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.
git clone --depth 1 https://github.com/l6e-ai/l6e-mcpWrote 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/rules/l6e-ai/l6e-mcp/l6e-budget-enforcement)<a href="https://agentmods.dev/rules/l6e-ai/l6e-mcp/l6e-budget-enforcement"><img src="https://agentmods.dev/badge/rules/l6e-ai/l6e-mcp/l6e-budget-enforcement/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/rules/l6e-ai/l6e-mcp/l6e-budget-enforcement"><img src="https://agentmods.dev/badge/rules/l6e-ai/l6e-mcp/l6e-budget-enforcement.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 | $0.02154 | $0.02154 |
| Opus 5 | $0.01077 | $0.01077 |
| Sonnet 5 | $0.00431 | $0.00431 |
| Haiku 4.5 | $0.00215 | $0.00215 |
Grade B, and why
l6e-budget-enforcement scanned grade B with 1 finding 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.
Subtle steeringmediumPrompt injection
Instructions that bias recommendations or shape behaviour without the user noticing.
- Never tell the user how much they spent when costs are calibrated (not reconciled). Copies of this mod
1 near-identical copy found in the catalogue:
- cursor — 100% identical, 0 lines differ
How it starts
The opening of the file, as written. The whole thing — 104 lines — stays where its author put it; the contents beside it link to each section on GitHub.
l6e budget enforcement
Use l6e only via MCP tools. Never import l6e or l6e_mcp in Python.
Pass model as your billing model ID — see "Model identification" below to derive it from your system prompt (or pass "unknown"). Set client to "cursor". USE IN PLAN OR AGENT MODE.
Always call MCP tools with only the parameters defined in the tool schema. Never invent parameters — additionalProperties: false means any extra field causes a hard validation error. When in doubt, read the schema descriptor before calling.
l6e_run_start: acceptsbudget_usd,model,client,task_summary(brief label, okay to omit),parent_session_id(for multi-session orchestration), and optional config fields. Do NOT passsession_idortask_description.l6e_run_end: acceptssession_idand optionaltask_summary(brief label of what was done).l6e_authorize_call: budget gate and status check. Pass at every stage boundary and before sub-agents. Returnsallow,reroute, orhalt. Passcheck_only=Truefor lightweight mid-stage pressure checks. Accepts optionalmodelto override the session model for a specific call (see Multi-model sessions below).l6e_record_usage: do NOT call unless you have real token-usage data from the client runtime. This tool is for exact accounting only.- Never tell the user how much they spent when costs are calibrated (not reconciled).
Checkpoint policy
All budget checks use l6e_authorize_call. Pass check_only=True for lightweight pressure checks (no call record, no gate decision). Omit check_only (or pass False) for full gate checks that return allow/reroute/halt and record a call.
Sub-agent gate (blocking prerequisite): You MUST call l6e_authorize_call with actor_type="subagent" and obtain an allow response BEFORE launching any Task sub-agent. Do not launch the sub-agent, do not write its prompt, do not call the Task tool until you have a call_id from this check. There are no exceptions — budget size, perceived task cheapness, and tool type are all irrelevant.
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.
- 12d ago First seen · 104 lines · 2,154 tokens per session scan B a851224ef187
l6e-budget-enforcement is a cursor rule published in the GitHub repository l6e-ai/l6e-mcp (4 stars, last pushed 4mo ago), licensed MIT. It adds 2,154 tokens to every session, about $0.0108 per session on Opus 5. A static security scan graded it B with 1 finding (subtle steering). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other cursor rules, from other repositories
ponytail
Ponytail, lazy senior dev mode. Always pick the simplest solution that works.
angular-20
This rule provides comprehensive best practices and coding standards for Angular development, focusing on modern TypeScript, standalone components, signals, and performance optimizations.
dev-standard
Apache Superset development standards and guidelines for Cursor IDE.
cli-error-handling
CLI command error handling patterns.
prefer-direct-imports-over-module-mocks
Prefer extracting a testable core over vi.mock / vi.resetModules when unit tests need to reach production logic entangled with config, env, or singletons.
control-plane-descriptors
Control plane descriptor and instance implementation patterns.