agentutility(7)$ man agentops
agentops
Trace, eval, and permission checks for autonomous work.
NAME
agentops — trace, eval, and permission checks for autonomous work
DESCRIPTION
Frontier probe for agents that already run multi-step work and need trace briefs, flaky-step detection, tool-call diffs, MCP permission review, red-team prompts, and cost-per-success scoring.
BUYERS — who calls into this cluster
Agent operators, MCP server builders, AI QA teams, and automation teams watching autonomous tasks fail in production.
WEDGE — why this exists vs. incumbents
Observability tools watch apps. AgentOps watches the paid agent workflow itself: tool calls, permissions, replay evidence, cost, and success criteria in one agent-callable surface.
ENDPOINTS — 4 entries · $0.008 – $0.02 per call
| # | endpoint | description | price | |
|---|---|---|---|---|
| 001 | agent-trace-brief | Turns a raw agent execution trace into a readable agent trace summary: an ordered step-by-step account of what the agent did, where it br… | $0.02 | → |
| 002 | eval-case-generate | Eval case generator for an agent or LLM task: describe the task and get back a draft llm eval dataset of {input, expected, rubric} cases… | $0.02 | → |
| 003 | mcp-tool-risk-score | Heuristic MCP tool risk review for an MCP tool manifest before you wire it into an agent. | $0.02 | → |
| 004 | tool-call-diff | Deterministic tool call diff between two agent run sequences. | $0.008 | → |
INSTALL AS MCP — add to any MCP-aware agent
npx -y @agentutility/mcp-agentopsFull Claude Desktop / Cursor / Cline configs → /mcp/agentops/
OTHER CLUSTERS — the portfolio compounds
| cluster | tagline | n | price |
|---|---|---|---|
| prooflayer | Trust + risk scanners for AI-built apps. | 17 | $0.005 – $0.25 |
| edgemarket | Onchain decisions agents can charge for. | 64 | $0.001 – $0.10 |
| edgefinance | TradFi signals at agent prices. | 89 | $0.002 – $0.05 |
| synthforge | Generative media for agents that ship products. | 59 | $0.01 – $0.50 |
| mediakit | One endpoint per format. Pay per call. | 67 | $0.003 – $0.20 |
| wordmint | Named LLM tasks. Minted at a sticker price. | 113 | $0.002 – $0.03 |
| webprobe | Inspect any website. Pay per look. | 155 | $0.001 – $0.10 |
| locale | Geo, weather, civic, identity primitives. | 40 | $0.002 – $0.05 |
| rollforge | The whole tabletop toolkit, one call at a time. | 26 | $0.003 – $0.08 |
| bestiary | A creature compendium for every world. | 6 | $0.01 – $0.08 |
| statline | Sports and stats math, per call. | 1 | $0.003 |
| matchpoint | Dedup and record linkage, per call. | 7 | $0.003 – $0.005 |
| retail | Product copy + product discovery, per call. | 8 | $0.01 – $0.02 |
| browserworkflow | Workflow memory for brittle browser and desktop agents. | 3 | $0.01 – $0.02 |
| modelrouter | Routing and cost checks for mixed model stacks. | 2 | $0.005 – $0.01 |
| compose | One call, several of our endpoints, one settlement. | 128 | $0.02 – $0.10 |
SEE ALSO