← Capabilities
LLM agents and tool use
not yet earnedI can choose between a workflow and an agent and justify it, sandbox what a model is allowed to touch, and measure multi-step reliability rather than assert it.
What this claim costs
- 5 assessed items at tier 2 or above0 of 5 — 5 more to do, from 21 available
- A public artifact built for this claim — agent-vs-workflowNot built yet
- A documented negative result — where this stops workingNothing recorded. A capability with no known failure boundary has not been explored.
- Defended aloud, against questions you did not write1 available, none done
- Recalled cold, a week later, without notes6 scheduled, none done
The proof
agent-vs-workflow
One stage of a working deterministic pipeline rebuilt as an autonomous agent, with the cost, latency and reliability of both measured side by side.
What it has to contain
- One working deterministic pipeline stage rebuilt as an autonomous agent
- Cost, latency and multi-step reliability measured for both
- The sandbox boundary stated: what the model may and may not touch
- A verdict on workflow versus agent, with the conditions that would flip it
Where it stops working. Not established yet. Anyone can ship a demo; the boundary is the part that is worth reading, so this claim is not earned until it is written down.
Skills underneath
- Planning and decomposition
- Orchestration patterns
- Multi-step reliability
- Sandboxing and least privilege
- Evaluating agents
- Multi-agent systems
21 assessed items across the specified courses evidence these, 0 done.