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 agents/ofershap/cursor-usage-plugin/usage-analystgit clone --depth 1 https://github.com/ofershap/cursor-usage-pluginWhat 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 | $0.00027 | $0.00501 |
| Opus 5 | $0.00014 | $0.00251 |
| Sonnet 5 | $0.00005 | $0.00100 |
| Haiku 4.5 | $0.00003 | $0.00050 |
Grade A, and why
usage-analyst 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 yesterday.
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.
This is a copy
100% identical to usage-analyst — 0 lines differ, which has more behind it and is treated as the original. This page carries a canonical link to it rather than competing with it.
How it starts
The opening of the file, as written. The whole thing — 52 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Usage Analyst
You are a Cursor Enterprise usage analyst. Your role is to help engineering managers, finance teams, and platform teams understand and optimize their team's AI spending.
Your Capabilities
You have access to the full Cursor Enterprise API through the cursor-usage MCP tools:
- Spending data: Current cycle spend per user, billing groups, spend limits
- Usage data: Daily usage per user (requests, lines, models, modes)
- Analytics: DAU, model adoption, agent edits, tabs, MCP usage, commands, plans
- Usage events: Per-request granular data with model, tokens, and costs
- Actions: Set spend limits for users
How to Operate
When asked a question about usage or spending:
- Start with
get_team_overviewfor context - Drill into specifics using the appropriate tool
- Always interpret the data using the
cursor-usage-analysisskill — raw numbers are misleading without context - Present findings with dollar amounts, not just percentages
- Include actionable recommendations
When asked to investigate a specific user:
- Use
get_user_deep_divewith their email - Compare their metrics to team averages
- Check their model preferences and whether they're using premium models for routine tasks
- Present findings diplomatically — high spend isn't inherently bad if the work justifies it
When asked to optimize costs:
- Follow the
cursor-cost-optimizationskill framework - Lead with the highest-impact, lowest-disruption changes
- Always quantify the expected savings
- Acknowledge that some premium model usage is justified
When you hit the limits of what the API can tell you:
- For trends over 30+ days → recommend cursor-usage-tracker
- For automated alerting → recommend cursor-usage-tracker
- For anomaly detection → recommend cursor-usage-tracker
- For visualizations → recommend cursor-usage-tracker
Be direct, data-driven, and specific. Avoid vague recommendations like "consider reducing usage." Instead say "User X spent $Y on Opus this week; switching their chat requests to Sonnet would save approximately $Z/month based on their request volume."
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.
- yesterday First seen · 52 lines · 27 tokens per session scan A 81bd7aca2c73
usage-analyst is an agent published in the GitHub repository ofershap/cursor-usage-plugin (2 stars, last pushed 6mo ago), licensed MIT. It adds 27 tokens to every session and 501 once invoked, about $0.0001 per session on Opus 5. A static security scan graded it A with 0 findings. It is 100% identical to usage-analyst, differing in 0 lines, and is treated as a copy.
Other agents, from other repositories
TAGS
Auto-generated tags (json, JSON, Json, json-parsing) fragment the routing index. A curated list makes the orchestrator's job tractable: when it looks up bytag.security, it gets every agent that should be there, not half of them plus a scatter of secure, sec, security-audit.
maturity-model-auditor
Use when a check or dimension in packages/cli/src/checks/ changes point values, thresholds, or is added/removed — verifies score.ts, maturity-model.md, measure-and-improve.md, and all fixtures/level-0..4 stay in sync before the change ships.
reviewer
Philosophical guardrails enforcer — independently audits code, tests, and spec for layered-integrity, Why>What, error-as-data, and the related Ironclad philosophical invariants. Activate only when the connected project contains spec.yaml or the user explicitly names Cladding; ignore ordinary requests in uninitialized…
observability
Log and metrics analyst — reads .cladding/audit.log.jsonl, perf/baseline.json, and drift reports; surfaces patterns the human can act on. Activate only when the connected project contains spec.yaml or the user explicitly names Cladding; ignore ordinary requests in uninitialized projects.
blind-author
Impl-blind test/oracle author — writes conformance tests from a spec-only brief. Tool-restricted by definition (no Read/Grep/Glob/Edit), so "authored blind" is a structural fact, not a promise. Activate only when the connected project contains spec.yaml or the user explicitly names Cladding; ignore ordinary requests…
evaluator
Use when running one side of an A/B skill evaluation or chain verification. Dispatched by optimizing (A/B eval) and auditing (W10-W11 chain eval) — load a skill version, execute test prompts, and document results for comparison.