Hermes Action Bridge
Use this skill when a task needs state or capabilities owned by Hermes Agent: Hermes skills or memory, connected services, messaging platforms, browser workflows, scheduled automation, or another integration configured in Hermes.
Do not reimplement a Hermes-owned integration in the coding agent. Prefer the
installed hermes_* MCP tools so the host can discover, approve, and inspect
the delegation.
When to use
- The task names Hermes, a Hermes skill, memory, preset, or connected service.
- The task needs messaging, scheduling, browser work, or persistent external automation Hermes owns.
- The task must continue as a cancellable background job outside the local coding step.
Do not use Hermes for ordinary local code edits or repository inspection that the current agent can perform directly.
Discover first
Call hermes_capabilities before assuming an integration is configured. Use
hermes_status to check the runtime without spending provider tokens.
Safe delegation
- Use
hermes_planfor plan-mode analysis, but treat it as open-world and review its output. - Use
hermes_submit, thenhermes_job_status/hermes_result, for a long cancellable plan or draft. - Use
hermes_cancelwhen a queued or running job is no longer wanted.
External side effects
For public posts, outbound messages, deletes, payments, credential changes, or git pushes:
- Call
hermes_prepareto produce a local no-tool summary and approval ID. - Call
hermes_approveto open the MCP client's interactive confirmation form. The bridge shows the exact action and executes only if the human accepts and checks the confirmation field. - Call
hermes_rejectwhen the action is no longer wanted.
Never answer or simulate the confirmation form yourself. Clients without MCP form elicitation fail closed; use plan, draft, or direct supervised CLI work instead of trying to bypass that boundary.
CLI fallback
If MCP tools are unavailable, use the local CLI:
hermes-action run --mode plan "<describe what Hermes should do>"
Verify setup
hermes-action doctor