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 agentmods add rules/elumenotion/guideants/local-ai-lifecycle-authoritygit clone --depth 1 https://github.com/Elumenotion/GuideAntsWrote 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/elumenotion/guideants/local-ai-lifecycle-authority)<a href="https://agentmods.dev/rules/elumenotion/guideants/local-ai-lifecycle-authority"><img src="https://agentmods.dev/badge/rules/elumenotion/guideants/local-ai-lifecycle-authority.svg" alt="Measured on agentmods" 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 | $0.00809 | $0.00809 |
| Opus 5 | $0.00404 | $0.00404 |
| Sonnet 5 | $0.00162 | $0.00162 |
| Haiku 4.5 | $0.00081 | $0.00081 |
Grade A, and why
local-ai-lifecycle-authority 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 6d 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 — 64 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Local AI lifecycle authority
GuideAntsApialone decides enabled/disabled state and model/bundle selection.ServiceModesis authoritative for auxiliary service routing and selections.- For local chat models, active quant / artifact selection for an alias is
owned by API installation provenance (
LocalModelInstallationsand the change-quant / install flows that write it). Router INI, loaded--modelargs, llama-admin journals, and files under the model store are apply outputs only — never a second source of “which quant is selected.” ga-adminonly validates and executes a complete plan explicitly submitted by the API. It must start empty/idle and must not load engines on boot.- Engine folders, inventory endpoints, status, and marker files are diagnostic outputs. Never use them to populate ServiceModes or choose what loads.
- Do not add a persisted desired-state plan or startup replay.
- Do not add auxiliary engine autoload environment flags or startup load calls.
- Do not add revision noop, runtime drift checks, or transition diffing in
warmup_orchestrator.py— that file is a dumb executor only; verification lives inLocalAiRuntimeAlignmentVerifier(GuideAntsApi). - Cloud-active routing must produce
enabled: falsefor the corresponding local engine and must skip load-specific setup such as SD bundle projection. - Missing API-owned local selections are errors. Do not guess from disk or engine state.
- Each local stack uses one holistic host URL:
LlamaCpp:BaseUrlfor chat,LocalServiceHosts:*BaseUrlfor aux. Lifecycle apply must target that same host (via per-stack plan split), not always the llama host.
Read docs/local-ai-lifecycle/ARCHITECTURE.md before changing local AI startup,
routing, warmup, model selection, engine admin, or compose behavior.
Debugging and incident investigation (same severity as design)
“Only diagnosing” is not an exception. Inspecting the AI container to learn what is selected is the same failure mode as autoloading from inventory.
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.
- 6d ago First seen · 64 lines · 809 tokens per session scan A 09673c8386a0
local-ai-lifecycle-authority is a cursor rule published in the GitHub repository Elumenotion/GuideAnts (57 stars, last pushed 2d ago), licensed Apache-2.0. It adds 809 tokens to every session, about $0.0040 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.
Other cursor rules, from other repositories
cursor-fable-rules
Fable coding lifecycle discipline and situational awareness for Cursor.
llm-router
When the user asks about architecture, system design, project structure, or "how should I build X", ALWAYS use the planworkflow MCP tool first.
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.