evolve-loop GEMINI.md

Project instructions for the Evolve Loop repository when it is used with Gemini CLI. They explain which files to read, where skills are found, and how Gemini routes runtime work through Claude in hybrid mode.

In plain words
What is it for?
Use them when developing Evolve Loop with Gemini CLI, discovering its skills, invoking the evolution loop, or maintaining its cross-agent runtime behavior.
Why use it?
They keep different coding agents aligned with the repository's shared rules and prevent incorrect skill or runtime setup.

Instructions file for Gemini CLI

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add instructions/mickeyyaya/evolve-loop/gemini-md
Clone the repo
git clone --depth 1 https://github.com/mickeyyaya/evolve-loop

Made for: Gemini CLI.

Per session 801 This file is loaded in full into every session.
When invoked 801 The same file — it is already loaded in full.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5 $0.00801 $0.00801
Opus 5 $0.00400 $0.00400
Sonnet 5 $0.00160 $0.00160
Haiku 4.5 $0.00080 $0.00080

Measured 2d ago against content hash 636b195acfe0, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

evolve-loop GEMINI.md scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

GEMINI.md · 61 lines

How it starts

The opening of the file, as written. The whole thing — 61 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Project Instructions (Gemini CLI)

Read AGENTS.md first — it carries the cross-CLI invariants and the 12 Core Agent Rules that bind every agent regardless of CLI. This file is the Gemini-specific overlay: skill discovery path, hybrid runtime, and tool name translation.

Release notes: CHANGELOG.md. Claude Code-specific runtime details: CLAUDE.md (not loaded into Gemini context — for cross-CLI maintainers only).

Skill discovery

Skills are auto-discovered at .agents/skills/<name>/SKILL.md (the cross-CLI open standard). The .agents/skills/ directory contains symlinks to the canonical skills/<name>/ location. Both paths resolve to the same SKILL.md content.

To invoke the primary skill: /evo:loop (registered via the plugin's slash-command set).

Runtime adapter (tier-1-hybrid)

.claude-plugin/plugin.json:compatibility.tiers declares gemini-cli: tier-1-hybrid:

  • Skill content is portable. The same SKILL.md works in Gemini CLI and Claude Code.
  • Runtime execution runs through the native Go binary (go/bin/evolve). In HYBRID mode Gemini routes runtime work to the claude binary while Gemini hosts the skill activation (per AGENTS.md). Gemini CLI lacks the non-interactive prompt mode (gemini -p) and subagent dispatch primitives the kernel hooks require for structural enforcement, so the runtime path delegates to claude rather than running natively. The shell shim lives at adapters/gemini.sh.

You need claude installed and authenticated for the runtime path. If only gemini is available, only skill text is usable — no autonomous cycle execution.

Tool name translation

Repository documentation uses Claude Code tool names. Gemini equivalents:

Claude Code Gemini CLI
Read ReadFile
Bash RunShell
Edit replace
Write write_file
Grep SearchCode
Glob SearchFiles
Skill activate_skill

Full table: skills/loop/reference/gemini-tools.md.

Read the full file on GitHub · 61 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 2d ago First seen · 61 lines · 801 tokens per session scan A 636b195acfe0

Subscribe to this mod's changes

evolve-loop GEMINI.md is an instructions file published in the GitHub repository mickeyyaya/evolve-loop (5 stars, last pushed 2d ago), licensed Apache-2.0. It adds 801 tokens to every session, about $0.0040 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.