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 jpantsjoha/ai-native-developer-experience --skill cost-guardrailgit clone --depth 1 https://github.com/jpantsjoha/ai-native-developer-experienceWrote 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/jpantsjoha/ai-native-developer-experience/cost-guardrail)<a href="https://agentmods.dev/skills/jpantsjoha/ai-native-developer-experience/cost-guardrail"><img src="https://agentmods.dev/badge/skills/jpantsjoha/ai-native-developer-experience/cost-guardrail/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/jpantsjoha/ai-native-developer-experience/cost-guardrail"><img src="https://agentmods.dev/badge/skills/jpantsjoha/ai-native-developer-experience/cost-guardrail.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 | $0.00055 | $0.01055 |
| Opus 5.5 | $0.00022 | $0.00422 |
| Sonnet 5.5 | $0.00011 | $0.00211 |
| Haiku 4.5 | $0.00006 | $0.00105 |
Grade A, and why
cost-guardrail 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 19d 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 — 86 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Cost Guardrail
The most expensive model is the one running on every request when it does not need to.
LLM cost is not a finance problem — it is an architecture problem. The design determines the bill. This skill enforces cost-awareness as a first-class design constraint, not an afterthought.
When to use
- Designing any system that calls an LLM (directly or via an agent)
- Before scaling a workload to higher volumes
- When a cost estimate is required for a feature or release
- When reviewing an architecture for unbounded cost vectors
- When choosing between model tiers for a given task
Procedure
-
Identify every LLM call in the system — list: which agent or component makes the call, the model tier used, the approximate input and output token counts, and the call frequency (per user action / per minute / per batch).
-
Apply the model tiering test — for each LLM call, ask:
- Does this task require deep reasoning, or is it classification / extraction / reformatting?
- Can the task be completed with a smaller or faster model?
- Is the model tier choice based on evidence (benchmark, A/B test) or assumption?
Starting hypotheses only (verify current pricing and qualify each task family before relying on them):
Task type Appropriate tier Simple classification, extraction, summarisation Small / fast model Complex reasoning, multi-step planning, code generation Mid-tier model Deep analysis, architecture decisions, adversarial review Highest-tier model A short ticket or one-file change does not qualify a cheap model. Require independent acceptance, preserve consequence floors, count failed attempts and bound escalation. Use the project's explicitly adopted routing mode; the routing guide is not automatic consumer activation. Model/effort changes require a fresh baseline.
-
Identify unbounded cost vectors — flag any call pattern where the token count or call volume has no upper bound:
- Loops that call an LLM until a condition is met (with no max-iteration guard)
- User-triggered calls with no rate limiting
- Context windows that grow unboundedly across a conversation
- Batch jobs with no per-run budget ceiling
What ships with it
1 file beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
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.
- 19d ago Changed · +5 lines 1dc808b67772
- 2mo ago First seen · 81 lines · 55 tokens per session scan A 7fc2123013ed
cost-guardrail is a skill published in the GitHub repository jpantsjoha/ai-native-developer-experience (13 stars, last pushed 19d ago), licensed Apache-2.0. It adds 55 tokens to every session and 1,055 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 2026-08-30.
Other skills, from other repositories
mlops-engineer
Provides MLOps patterns for ML CI/CD pipelines, model registries, monitoring, and data drift detection. Use when setting up ML infrastructure or when the user mentions MLOps, model deployment, ML pipeline, or model monitoring.
agent-platform-tuning
Agent Platform Model Tuning. Use when you need to fine-tune open models or Gemini models using Agent Platform infrastructure. Don't use for model training outside Agent Platform, model deployment to endpoints (use agent-platform-deploy), or managing serving endpoints (use agent-platform-endpoint-management).
google-cloud-solution-guided-gke-ai-migration
Guides the migration of existing AI workloads (Cloud Run, Gemini API, Gemini Enterprise Agent Platform) to self-hosted GKE inference using gcloud and kubectl. Use when the user has an existing AI inference workload (on Cloud Run, the Gemini API, Gemini Enterprise Agent Platform, or a custom VM) and wants to move it to…
apigee-agent-gateway-governance
Stands up an Apigee API proxy that a Google Cloud Agent Gateway calls as an ext-proc (Service Extension) callout on an agent's LLM traffic, to meter token usage per model, extract caller identity, and apply Model Armor prompt screening or a semantic cache. Use when wiring Apigee in as an AI gateway behind an Agent…
agent-platform-endpoint-management
Manages Agent Platform serving endpoints. Use when you need to create, list, describe, update, or delete serving endpoints for model deployment on Agent Platform. Also use when troubleshooting endpoint permission, quota, or resource busy errors. Don't use for deploying models to endpoints or for running model…
gke-inference
Deploys and optimizes AI/ML inference workloads on GKE, using GPUs, TPUs, and model servers. Use when deploying GKE inference servers, configuring GKE GPU resources for inference, or deploying LLMs on GKE. Don't use for migrating existing AI workloads to GKE (use google-cloud-solution-guided-gke-ai-migration), GKE RAG…