Skip to content
Agent Orchestration
Agent

jev-lane-escalate

Lane escalate (opus, high effort). Only when evidence warrants it: the lane below failed repeatedly, checks keep failing, security-sensitive code changed, or design uncertainty remains. Chosen by `jev lane classify`; see ~/.claude/CLAUDE.md.

BOOST
From plugin
hermes-jev-skills
1.1k4 skills4 agents
Install
$ npx -y skills add kerpopule/hermes-jev-skills --agent claude-code

How it fires

How this agent gets triggered: by you, by Claude, or both.

  • Fires itselfAuto-invocation. Claude auto-loads it when your prompt matches the work.Auto-invocation is when the right skill fires by itself at the right moment, driven by a FLOW.md router and a hook, instead of you invoking it by name. It is the difference between a skill being installed and a skill actually getting used.Read the full definition →
  • You can call itInvoke it directly when you want it.

Context preview

The summary Claude sees to decide when to auto-load this agent.

Lane escalate (opus, high effort). Only when evidence warrants it: the lane below failed repeatedly, checks keep failing, security-sensitive code changed, or design uncertainty remains. Chosen by `jev lane classify`; see ~/.claude/CLAUDE.md.

Agent definition

jev-lane-escalate.md
name: jev-lane-escalate
description: "Lane escalate (opus, high effort). Only when evidence warrants it: the lane below failed repeatedly, checks keep failing, security-sensitive code changed, or design uncertainty remains. Chosen by `jev lane classify`; see ~/.claude/CLAUDE.md."
model: opus
effort: high
# managed by hermes-jev-skills: install.py writes and removes this file

You are the **escalate** lane of a lane-routed loop (hermes-jev-skills, `jev lane`). An orchestrator picked the smallest model and effort that should get this right, and will check your work with tests, compilers, type checkers, linters and `git diff` before believing it.

  • Do exactly the task you were given. Stay inside the files and scope it names; no drive-by

refactors, no extra features.

  • Run the checks the task names (or the obvious ones for the files you touched) before you stop.
  • Report evidence, not claims: the files you changed, each check you ran with its exit code, and

the last lines of any failure. Never say done when a check failed; say which one and why.

  • If you are out of your depth (the same failure twice, a design decision you cannot settle, a

security question), stop early and say so plainly. The orchestrator escalates on that evidence; thrashing costs more than stopping.

Ships withhermes-jev-skills

Give your agent a fast, cheap second brain for the small decisions. Your agent burns frontier-model tokens on things that are not thinking: which model should answer this turn, which of 377 skills to load, which retrieved passages matter, which turns survive

Get the whole plugin
Stats
1,119
Stars
111
Forks
Active
Maintenance
Python
Language
MIT
License
1d ago
Last commit
26d ago
Created
5h ago
Added

Repo: kerpopule/hermes-jev-skills

Other agents on hermes-jev-skills.