Claude Code vs Hermes Agent: Scripts or Profile Handoffs?
Compare Claude Code and Hermes Agent on executable workflows, approval learning and completion checks for recurring development work.
Founder, Task Machine
Should Claude Code or Hermes Agent coordinate your recurring development work through an executable script or a board of specialist profiles? Claude Code documents JavaScript workflows that control loops, branches and agent stages. Hermes Agent describes dependency-linked Kanban cards, while its kernel leaves workflow-template routing for a future version.
That distinction matters when a maintenance job needs the same routing on every run. A board can show which specialist should work next, but a published dependency chain and an executable branching process make different promises. Start by deciding whether you need visible handoffs or a script that determines the next stage.
Claude Code makes orchestration readable as code
Claude Code's dynamic workflow documentation describes a script you can inspect and rerun. For a repository maintenance process, the script can express the order of agent stages and what happens when a branch or loop changes the route. The orchestration lives in a JavaScript artifact that your team can read.
There is also a less scripted coordination surface. Its Projects documentation describes a coordinator that starts threads, tracks their results and gathers threads requiring an answer under Waiting on you. Projects are public beta for Pro, Max and Ultra, with a gradual rollout, and are not yet offered on Team or Enterprise.
That availability boundary belongs in your evaluation. A developer testing a personal Project and a team purchasing a shared coding setup should check the specific surface they intend to use. Cloud project threads also do not apply repository permission rules, so a local configuration is not evidence that a cloud thread follows those same rules.
Hermes Agent centers the specialist profile
Hermes Agent's Kanban guide describes an orchestrator profile that owns a root task, delegates child tasks and returns to judge completion. Cards preserve comments, dependencies and attempt history. This gives a developer a visible record of which profile tried a job and which dependency still blocks it.
People can manage and comment on cards, while the documented assignee is an agent profile. The board uses a trusted local-user model and is single-host by design. That is a concrete fit for an installation you operate, but a boundary to examine before treating it as a shared work system across people.
The workflow distinction is narrower than whether Hermes Agent can coordinate agents. It can coordinate ordered handoffs, but its documented kernel ignores workflow-template routing fields. If a specification depends on those fields choosing a route, the public documentation does not establish that execution behavior.
Recurrence does not settle the ownership question
Claude Code's scheduled task guide covers recurring schedules, HTTP API calls and GitHub events. These are research-preview cloud runs, each with a separate session and history reachable from the saved task. Saved tasks belong to the individual account and use that account's connected identities, without sharing the task with teammates.
Hermes Agent cron jobs support one-off and recurring schedules, and its webhooks can trigger the same jobs immediately. A job can specify instructions, skills and delivery settings. For a self-hosted installation, the gateway must remain running for agent-backed jobs, and scheduler ticks skip those jobs when it is down.
For an overnight check, ask who owns the saved configuration and what must remain available. Claude Code's published route is an individual account's cloud task. Hermes Agent's self-hosted route makes gateway operation part of your maintenance responsibility. Neither question is answered by showing that a scheduled prompt ran once.
Permission learning and completion need separate tests
Hermes Agent's security guide describes shell approval modes, allowlists and deny rules. Its approvals suggest command turns previously executed, approved commands into ranked allowlist proposals. You choose which entries to apply, and destructive command classes are excluded.
Claude Code's auto-mode setup drafts environment and rule entries using project configuration and recent commands for your acceptance. Its public pages do not describe permission-change proposals learned from approval history. If your objective is to reduce repeated approval prompts, test the actual proposal each product produces, including the evidence it uses.
Their completion checks also inspect different evidence. Claude Code's /goal evaluator checks the stated completion condition after each turn using conversation evidence. It does not independently execute commands or read files. Hermes Agent's goal contracts define verification criteria and require shell quality gates to pass before completion.
A developer evaluating test repair should inspect how the reported test result reaches the completion decision. A declaration that the agent finished and a gate requiring a command to pass are different pieces of evidence.
Claude Code and Hermes Agent, point by point
The table gives both products the same questions, with the source and check date attached to each answer. Use it to distinguish the documented capability from the scope of its implementation.
Both columns answer the same questions from each company's public pages. Yes means the company publishes the capability the question asks about, Partly means something narrower, No means it says it does not offer it, and Not published means its public pages do not describe it. Every answer links to its source and shows when we last checked it. The oldest check on this page was November 13, 2026.
| What you compare | Claude Code | Hermes Agent |
|---|---|---|
| Goals |
Claude Code
Yes
The /goal command gives Claude a measurable completion condition, such as passing tests or an empty issue backlog. A separate model checks the condition after each turn and tells Claude to continue or stops the session's goal. |
Hermes Agent
Partly
Persistent goals keep an agent working toward an objective, with a judge and optional completion tests. Its docs describe no linked metric tracking against an ongoing business goal. |
| Leads |
Claude Code
Yes
Projects give Claude a coordinator role that starts and tracks threads, while experimental agent teams have a lead that assigns tasks and synthesizes results. Projects are public beta for Pro, Max and Ultra, with a gradual rollout. |
Hermes Agent
Yes
You can name an orchestrator profile to own a decomposed project's root task, assign child tasks to specialists, judge completion and add more work when needed. |
| Tasks |
Claude Code
Partly
Experimental agent teams share tasks with agent owners, dependencies and completion states, and Projects retain work in threads. Its docs do not describe the same work item being owned interchangeably by a person or an agent. |
Hermes Agent
Partly
Kanban cards keep an assigned agent profile, comments, dependencies and attempt history. People can comment and manage cards, but the documented assignee is a profile name rather than a person or an agent. |
| Recurring work |
Claude Code
Yes
Research-preview scheduled tasks run from recurring schedules, HTTP API calls or GitHub events. Each run opens a separate cloud session, with its history available from the saved task. |
Hermes Agent
Yes
Cron jobs run once or recur on a schedule, and webhook events can fire the same jobs immediately. Each job can carry its own instructions, skills and delivery settings. |
| Workflows |
Claude Code
Yes
Dynamic workflows put loops, branches and agent stages in a JavaScript script that you can read and rerun. The script controls what runs next rather than leaving each orchestration decision to the conversation. |
Hermes Agent
Partly
Kanban dependency chains coordinate roles such as scout, editor and writer. They provide ordered handoffs, but the kernel does not execute workflow templates, whose routing fields are reserved for a future version. |
| Autonomy |
Claude Code
Yes
Local permission rules allow, ask or deny tool actions, with approval prompts in the session. Public-beta Projects list threads needing approval under Waiting on you, although cloud threads do not apply repository permission rules. |
Hermes Agent
Yes
You configure shell approval modes, permanent allowlists and deny rules. Dangerous-command requests appear in the terminal or messaging chat, where you can approve or deny them. |
| Learned autonomy |
Claude Code
Partly
The /auto-mode-setup command drafts environment and rule entries from project configuration and commands in recent sessions, for you to accept. Its docs do not describe learning permission-change proposals from your approval history. |
Hermes Agent
Yes
The approvals suggest command mines executed, approved commands into ranked allowlist proposals. You choose which entries to apply, and destructive command classes are excluded. |
| Memory |
Claude Code
Yes
Auto memory keeps corrections and preferences across local sessions in files you can read, edit or delete. Public-beta Projects also have editable project memory that later cloud threads read. |
Hermes Agent
Yes
Persistent notes and user preferences carry into later sessions. You can inspect, edit or delete learned memory through the journey commands or the desktop panel. |
| Planning |
Claude Code
Yes
Plan mode researches a request and writes a proposed plan before implementation. You can revise the plan in your editor or ask Claude to keep planning before approving execution. |
Hermes Agent
Yes
From chat, Kanban can expand a rough idea into a specification or dependency-linked tasks. You can disable automatic child promotion so the generated tasks wait for manual review before execution. |
| Intake |
Claude Code
Yes
An HTTP POST can send an incoming alert to a saved Claude Code task whose prompt tells the agent how to investigate it. GitHub events can also start these research-preview cloud runs. |
Hermes Agent
Yes
Signed webhooks turn outside events into agent prompts. A route can target a named profile when profile multiplexing is enabled, and filter which events start work. |
| Connectors |
Claude Code
Yes
MCP servers connect external tools and APIs. Research-preview scheduled tasks accept HTTP API triggers and GitHub events, and let you select which connectors each run can use. |
Hermes Agent
Yes
A curated Model Context Protocol catalog and custom local or remote servers connect external tools. HTTP webhook routes accept signed service events and can post responses back through integrations such as GitHub. |
| Models |
Claude Code
Yes
The opusplan setting automatically uses Opus for planning and Sonnet for execution. Code Review also chooses its own Claude models rather than asking administrators to select one. |
Hermes Agent
Partly
You choose the main model and can override auxiliary tasks. Provider routing chooses a backend for a selected model, rather than selecting a model from each task's needs. |
| Cloud or your own computer |
Claude Code
Yes
Claude Code runs locally in your terminal and editors or in cloud sessions. Projects can send a thread to your computer through Remote Control when it needs local tools or files. |
Hermes Agent
Yes
Run the free agent locally through its desktop or terminal interfaces, or choose the optional Hermes Cloud service for a hosted instance with a persistent workspace. |
| Your own subscription |
Claude Code
Yes
You can sign in with a Claude subscription, supply an Anthropic API key or configure supported cloud-provider credentials. The published connections cover Claude usage rather than ChatGPT or GitHub Copilot subscription allowances. |
Hermes Agent
Yes
Hermes accepts your provider API keys and a ChatGPT account login. Its Claude account path requires Max plus purchased extra-usage credits, not the included Max allowance. |
| Pricing |
Claude Code
Yes
Claude Code supports token-based API billing and subscriptions with usage limits. Pro costs $20 billed monthly or $200 billed annually, Max $100 per month and Ultra $200 per month, excluding applicable tax. |
Hermes Agent
Yes
The MIT agent is free. Portal lists Plus $20/month, Super $100/month and Ultra $200/month, with credit top-ups. Cloud's FAQ separately meters compute and storage, excluding inference and tools, without explaining how that relates to included plan hosting. |
| Correctness |
Claude Code
Yes
The /goal evaluator checks a stated completion condition after each turn, such as tests passing, and sends unmet work back. It judges the conversation evidence and does not independently run commands or read files. |
Hermes Agent
Yes
Goal completion contracts define verification criteria, and shell quality gates must pass before a goal completes. Kanban review also maps acceptance criteria to evidence and can return work for correction. |
For software work
| What you compare | Claude Code | Hermes Agent |
|---|---|---|
| Code review rules |
Claude Code
Yes
Code Review flags changes that violate CLAUDE.md instructions and accepts review-specific rules in REVIEW.md. The hosted service is available on Team and Enterprise plans. |
Hermes Agent
Yes
The pull-request review guide uses reusable review guidelines and remembered team conventions, such as required type annotations and database access rules, to check later diffs. |
| Adversarial review |
Claude Code
Yes
Research-preview ultrareview launches a separate fleet of reviewers in a cloud sandbox, with findings independently reproduced and verified. Dynamic workflows can also have independent agents adversarially review each other's findings. |
Hermes Agent
Yes
A bundled pre-commit skill delegates security and logic review to a separate, diff-only reviewer in interactive sessions. Kanban's review skill independently verifies handoffs and keeps the reviewer from editing the implementation. |
| Risk rating |
Claude Code
Not published
Its docs describe no overall risk rating for a revision before merge. Code Review ranks individual findings by severity and leaves the merge decision with your team. |
Hermes Agent
Not published
Its docs describe no change-level risk rating before merge. Optional residual-risk notes on Kanban handoffs and severity labels for individual review findings do not establish such a rating. |
| Pull requests |
Claude Code
Yes
Public-beta Projects open pull requests from cloud threads and watch them with auto-fix. Threads push fixes for failed checks and address review comments before reporting the pull request ready for you. |
Hermes Agent
Partly
Agents can open pull requests through the GitHub CLI, and the bundled GitHub skill requires reading PR comments. The cited docs do not establish a review-feedback revision cycle. |
A fact we cannot find or cannot check is left out, and prices can change after the date shown. Task Machine publishes this comparison and is an interested party. Claude Code and Hermes Agent are named only to identify them, and there is no relationship between the companies. Has something changed? Tell us.
Who each one suits
Pick Claude Code if you want rerunnable JavaScript orchestration alongside coding sessions and cloud coordination. Its limit here is that Projects remain a gradually rolling public beta, and scheduled cloud tasks are tied to individual accounts.
Pick Hermes Agent if named specialist profiles, a self-hosted board and shell gates match how you operate. Its honest limit is that ordered card dependencies do not make the reserved workflow-template routing executable. Its single-host board also deserves attention before you extend ownership beyond your installation.
Where Task Machine fits
Task Machine makes a repeatable process a Workflow with branches, human questions, approvals and verifier checks that return weak work with feedback. Its Workflow builder documentation describes that execution boundary. Task Machine is the wrong pick if you need to import your Hermes Agent board, because its docs do not describe importing Hermes boards, session histories or skill files. See Switch from Hermes Agent to Task Machine.
For a process that crosses a developer and an agent, compare the ownership boundaries in Task Machine vs Claude Code and Task Machine vs Hermes Agent. Take one existing maintenance job and write down its required approval and retry branch, then check that exact route against the documented workflow in each comparison.