Frequently asked questions
Answers about Hermes Plant products, agent safety tooling, billing, and policies. Each question links to a dedicated page with more detail.
Looking for the finance & quant agent services — how AI agents pay per call over x402, what calculations are available, and pricing?
See the agent services FAQ →Finance & quant agent services (x402)
Can LLMs reliably compute financial math?
No. Language models predict tokens rather than running a numerical solver, so calculations like IRR, XIRR, Black-Scholes Greeks, or a yield-to-maturity come bac
Do these finance APIs need a market-data feed or API key?
No. The Hermes Plant finance endpoints are pure calculations over the inputs your agent already has — cashflows, holdings, option parameters, bond terms. There
How do AI agents pay for APIs?
With x402, an agent's first request returns HTTP 402 with a payment challenge. The agent's x402 client signs a small USDC-on-Base authorization and resends the
How much do pay-per-call finance APIs cost?
Hermes Plant's agent-service endpoints cost between $0.02 and $0.30 per call, settled in USDC on Base over x402. There is no subscription, minimum, or monthly f
What is a DCF API?
A DCF API computes a discounted-cash-flow valuation from a projected cashflow series and a discount rate, returning present value plus related metrics. Hermes P
What is the x402 protocol?
x402 is an open payment protocol built on the HTTP 402 'Payment Required' status code. A server responds to an unpaid request with a 402 challenge describing th
IncidentScribe
How does IncidentScribe parse agent JSONL logs?
IncidentScribe reads JSONL line by line, tolerantly skipping or warning on unknown records while extracting shell commands, file edits, and tool calls. Events a
How does IncidentScribe flag critical commands in a timeline?
Severity heuristics match destructive patterns in shell and SQL text — database drops, force pushes, recursive deletes, infrastructure teardown, and similar age
What is the difference between free IncidentScribe and the Postmortem Pack?
The free tier includes the browser timeline viewer, Markdown/HTML/JSON export, and the Python CLI with `render` and `validate` commands. The Postmortem Pack ($1
Does IncidentScribe upload my session logs to a server?
The free web timeline viewer processes JSONL in your browser — paste or upload stays client-side unless you choose to export. For air-gapped or sensitive sessio
How does the Postmortem Pack GitHub Action attach timelines to issues?
The `attach-timeline` composite action installs IncidentScribe via pip, runs `incidentscribe render` on a JSONL path you provide, uploads the output as a workfl
What files are included in the IncidentScribe Postmortem Pack?
The pack includes INSTALL.md, TEAM-POSTMORTEM-GUIDE.md (two-week rollout), FORENSIC-CHECKLIST.md, three templates (blameless, exec-summary, security-incident),
Which Postmortem Pack template should I use for an agent-caused incident?
Use blameless for standard engineering postmortems with action items — it is the default for internal incidents. Use exec-summary when stakeholders need duratio
ReviewQueue
Can AI agents bypass ReviewQueue?
Yes — ReviewQueue only intercepts commands routed through `reviewqueue exec --`. An agent with unrestricted shell access can run `/bin/rm` directly and skip the
How does reviewqueue exec decide what to queue?
The risk classifier inspects argv before subprocess execution. Patterns like git push, recursive deletes, and deploy commands are flagged as risky and written t
What is the difference between free ReviewQueue and the Team Pack?
The free tier includes the CLI (`init`, `exec`, `list`, `approve`, `reject`, `serve`, `export`), JSONL audit logs, and the web queue/audit viewers. The Team Pac
How do multiple reviewers share one ReviewQueue approval inbox?
Store `~/.reviewqueue` in a private Git repository: `reviewqueue init --queue-dir .`, commit `pending/` and `audit.jsonl`, and clone to each teammate machine. A
How do I generate a monthly ReviewQueue audit rollup with the Team Pack?
The pack includes `templates/audit-rollup.sh`, a shell wrapper around `reviewqueue export --days 30 --format markdown`. Set `REVIEWQUEUE_DIR` if your queue live
How do I connect ReviewQueue to Slack with the Team Pack?
Create a Slack Incoming Webhook in your workspace, export `SLACK_WEBHOOK_URL` on the agent or notify host, and copy `scripts/notify_slack.py` from the Team Pack
Why does the web app use paste instead of reading my local queue?
The Vercel-hosted app has no access to your laptop's `~/.reviewqueue/` directory. Use `reviewqueue serve` on localhost and connect a local dashboard, or copy/pa
What files are included in the ReviewQueue Team Pack?
The pack includes README.md, PUBLISH.md (maintainers), `scripts/notify_slack.py`, `scripts/notify_telegram.py`, `guides/shared-queue-git-sync.md`, `guides/team-
What is ReviewQueue?
ReviewQueue is a human-in-the-loop approval gate for risky AI-agent shell commands. The free CLI wraps `exec`, classifies command risk, queues dangerous command
Operator OS Core
Do I need Notion to use Operator OS Core?
No. The Notion template is optional convenience. You can use the exact column structures from templates/revenue-command-center.md in a Google Sheet, Apple Numbe
How does the Weekly Reconciliation prompt find revenue leaks?
You paste two lists: what you expected to receive (recurring MRR + due invoices) and what actually landed in Stripe, PayPal, or bank. Prompt 06 returns reconcil
How long does the Operator OS revenue week actually take?
The core loop (MRR Pulse + Client Health Scan + Reconciliation + Weekly Revenue Review) takes about 30 minutes once the dashboard has a few weeks of history. Fi
What are the eight prompts in Operator OS Core?
The eight are: 00 The Revenue Week (chaining guide), 01 Time & Attention Audit (bi-weekly), 02 MRR Pulse (weekly heartbeat), 03 Client Health Scan (weekly), 04
What is Operator OS Core?
Operator OS Core is a 30-minute weekly operating system of 8 AI prompts plus a Revenue Command Center dashboard (Notion or sheet) that turns scattered revenue w
General
What should teams capture before and after an AI agent incident?
Before a session: note agent and model version, branch, deploy environment, blast radius, and who approves destructive commands. During: watch for production sh
How do I use the findings from MCP Risk Analyzer?
Each finding cites the precise tool name, the matched factor, a why sentence explaining the blast radius, and a fix string you can apply directly (e.g. constrai
Can I run MCP Risk Analyzer locally without paying?
Yes. The MCP Risk Analyzer Operator Pack ships a free local CLI (mcp-risk score manifest.json --json) that executes the identical six-factor ruleset against fil
What is MCP Risk Analyzer?
MCP Risk Analyzer scores an MCP server manifest for security risk before an agent installs it. It returns a structured findings array instead of a single opaque
Which risk factors does MCP Risk Analyzer check?
The analyzer evaluates six factors on every tool in the manifest: destructive-action capability (delete, drop, exec, shell), over-broad-scope or missing least-p
Still need help?
For billing, delivery, or policy questions not covered here, contact support.
Contact support