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 skills/xiaowen-jiang/agent-enterprise/workforce_optimizationnpx skills add Xiaowen-Jiang/agent-enterprise --skill workforce_optimizationgit clone --depth 1 https://github.com/Xiaowen-Jiang/agent-enterpriseWrote 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/xiaowen-jiang/agent-enterprise/workforce_optimization)<a href="https://agentmods.dev/skills/xiaowen-jiang/agent-enterprise/workforce_optimization"><img src="https://agentmods.dev/badge/skills/xiaowen-jiang/agent-enterprise/workforce_optimization.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 | $0.00016 | $0.00551 |
| Opus 5 | $0.00008 | $0.00275 |
| Sonnet 5 | $0.00003 | $0.00110 |
| Haiku 4.5 | $0.00002 | $0.00055 |
Grade A, and why
Workforce Optimization 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 3d 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 — 68 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Workforce Optimization
You are conducting a workforce cost and performance optimization review.
Process
-
Gather Data: Call
analyze_workforceto get the current workforce snapshot (costs, performance scores, model assignments, provider costs). -
Analyze: Review each agent against the decision matrix below.
-
Recommend: Produce a prioritized list of changes with expected impact.
-
Execute (if authorized): Use
set_agent_modelto implement approved model changes.
Decision Matrix
| Performance | Cost | Action |
|---|---|---|
| High | High | Consider cheaper model — agent delivers well, may not need expensive model |
| Low | High | Investigate — may need termination, retraining, or role change |
| High | Low | Keep current setup — optimal cost-performance ratio |
| Low | Low | May need a more capable model or clearer task instructions |
Cost Optimization Rules
- Workers doing routine tasks (coding, research) should use cheaper models (Haiku, GPT-4o-mini, Gemini Flash)
- Managers doing synthesis and delegation need capable models (Sonnet, GPT-4o) — do not downgrade without CEO approval
- Check that the target model's provider has a configured API key before recommending
- Consider the agent's task history — complex tasks may justify a more expensive model
Budget Right-Sizing
- Agents with >50% budget remaining after significant task volume may have over-allocated budgets
- Agents near their limit with good performance may need budget increases
- Calculate cost-per-completed-task when possible to compare efficiency across agents
Output Format
### Workforce Optimization Report
**Analysis Date**: [date]
**Total Workforce Cost**: $X.XX
#### Recommendations
1. [PRIORITY: HIGH/MED/LOW] Agent: [name] — Action: [description] — Rationale: [why]
2. ...
#### Summary
- Agents analyzed: N
- Changes recommended: N
- Primary optimization: [model right-sizing / budget adjustment / restructuring]
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.
- 3d ago First seen · 68 lines · 16 tokens per session scan A 37f89e491ece
Workforce Optimization is a skill published in the GitHub repository Xiaowen-Jiang/agent-enterprise (52 stars, last pushed 5mo ago), licensed MIT. It adds 16 tokens to every session and 551 once invoked, about $0.0001 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 skills, from other repositories
multi-model-router
多模型智能路由 - 根据任务复杂度评分选择最优模型，平衡成本与质量.
cost-tracker
Track, estimate, and optimize LLM API costs across providers. Use when user wants to understand their LLM spending, reduce costs, or compare provider pricing for a given workload.
sql-sentinel
Audit SQL for the cost & performance anti-patterns that burn warehouse credits. Catches SELECT , full-table scans, non-sargable predicates, Cartesian joins, NULL-trap NOT IN, and 17 more rules. Scores warehouse health 0-100 and outputs a prioritized cost-reduction plan for BigQuery, Snowflake, Redshift, and Postgres.
data-cost-optimization
Use this skill when asked about data cost optimization, Snowflake cost management, BigQuery slot management, S3 storage tiering, warehouse cost optimization, FinOps for data, cloud data cost reduction, or query cost analysis. This skill enforces: Snowflake credit usage analysis with query optimization, BigQuery slot…
gptq
Post-training 4-bit quantization for LLMs with minimal accuracy loss. Use for deploying large models (70B, 405B) on consumer GPUs, when you need 4× memory reduction with <2% perplexity degradation, or for faster inference (3-4× speedup) vs FP16. Integrates with transformers and PEFT for QLoRA fine-tuning.
ml-training-recipes
Battle-tested PyTorch training recipes for all domains — LLMs, vision, diffusion, medical imaging, protein/drug discovery, spatial omics, genomics. Covers training loops, optimizer selection (AdamW, Muon), LR scheduling, mixed precision, debugging, and systematic experimentation. Use when training or fine-tuning…