What is agent-guard GEMINI.md?

A local, read-only audit extension that checks past AI coding sessions and assistant configuration for exposed secrets or other security findings. It masks credentials and does not make network calls or change files.

Use it to audit a session before sharing a transcript or repository, review local assistant settings, list session metadata, and view aggregate finding counts.

It lets you inspect possible leaks without revealing the masked values or modifying the audited machine.

agent-guard GEMINI.md — Instructions file for Gemini CLI by skarn-security: A local, read-only audit extension that checks past AI…

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/skarn-security/agent-guard/gemini-md

Clone the repo

git clone --depth 1 https://github.com/skarn-security/agent-guard

Made for: Gemini CLI.

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

README.md

[![agentmods](https://agentmods.dev/badge/instructions/skarn-security/agent-guard/gemini-md.svg)](https://agentmods.dev/instructions/skarn-security/agent-guard/gemini-md)

Your own site

<a href="https://agentmods.dev/instructions/skarn-security/agent-guard/gemini-md"><img src="https://agentmods.dev/badge/instructions/skarn-security/agent-guard/gemini-md.svg" alt="Measured on agentmods" height="20"></a>

Per session 320 This file is loaded in full into every session.

When invoked 320 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.

Model	Per session	Once invoked
Fable 5	$0.00320	$0.00320
Opus 5	$0.00160	$0.00160
Sonnet 5	$0.00064	$0.00064
Haiku 4.5	$0.00032	$0.00032

Measured 4d ago against content hash 58f13f02dab3, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

agent-guard 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 4d 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 · 19 lines

What it actually says

Skarn

This extension adds a local skarn MCP server and the skarn-audit skill. The server runs on the user's machine, makes no network call, and exposes four tools:

scan_sessions returns findings from past AI coding sessions. Every previewed value is redacted.
vet_configs returns a report on the local assistant configuration surface. Every value is masked.
list_sessions returns session metadata: identifiers, timestamps, projects, counts. Never message content.
session_stats returns aggregate counts. Never message content.

Every tool is read-only. None of them writes a file, changes a configuration, or reaches the network.

A masked value is masked because it is a credential. Never reconstruct one, never infer the original from the surrounding text, and never ask the user to paste the unmasked value back. Report the finding by its rule, severity, and location instead.

Reach for the skarn-audit skill when the user asks whether a session or a configuration leaked something, or asks for an audit before sharing a repository or a transcript. This extension carries that skill only.

The guard hook in this extension runs in audit mode. It reports the verdict it would have reached and never blocks what the agent does.

If gemini mcp list shows the skarn server as Disconnected, run skarn --version in the same shell. command not found means the binary is not installed or is not on the PATH the CLI inherits. Install it as the repository README describes.

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.

4d ago First seen · 19 lines · 320 tokens per session scan A 58f13f02dab3

Subscribe to this mod's changes

agent-guard GEMINI.md is an instructions file published in the GitHub repository skarn-security/agent-guard (33 stars, last pushed 3d ago), licensed MIT. It adds 320 tokens to every session, about $0.0016 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 2026-08-30.

Security gemini-cli-extension

agent-guard GEMINI.md

Getting it into your agent

Wrote this? Show the measurements

What it costs to keep this loaded

Grade A, and why

Nothing flagged

What it actually says

Skarn

What this file has done since we first saw it

Other instructions, from other repositories

cli AGENTS.md

notfair-plugin AGENTS.md

Agentkey AGENTS.md

domscribe CLAUDE.md

adeu GEMINI.md

bookstack-mcp CLAUDE.md