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.
git clone --depth 1 https://github.com/dshakes/compassWrote 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.
[](https://agentmods.dev/plugins/dshakes/compass/gemini-extension)<a href="https://agentmods.dev/plugins/dshakes/compass/gemini-extension"><img src="https://agentmods.dev/badge/plugins/dshakes/compass/gemini-extension/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/plugins/dshakes/compass/gemini-extension"><img src="https://agentmods.dev/badge/plugins/dshakes/compass/gemini-extension.svg" alt="Reviewed on agentmods" width="80" height="20"></a>Grade A, and why
compass 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 1mo 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.
What it actually says
{
"name": "compass",
"version": "2.2.0",
"description": "compass — one operating manual + eval-gated guardrails, red-team hardening, a cost-tier router and signed config for your coding agent. Loads the manual as GEMINI.md and wires the same version-pinned MCP servers (context7 · fetch · git) compass uses for Claude Code and Codex.",
"contextFileName": "GEMINI.md",
"mcpServers": {
"context7": {
"command": "npx",
"args": [
"-y",
"@upstash/[email protected]"
]
},
"fetch": {
"command": "uvx",
"args": [
"--from",
"mcp-server-fetch==2025.4.7",
"mcp-server-fetch"
]
},
"git": {
"command": "uvx",
"args": [
"--from",
"mcp-server-git==2026.1.14",
"mcp-server-git"
]
}
}
}
What ships with it
60 files beside gemini-extension.json in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
- assets/ablation.svg 12 KB
- assets/budget-real.gif 3258 KB
- assets/budget.gif 276 KB
- assets/compass-rose.svg 1.2 KB
- assets/explainer.svg 10 KB
- assets/fleet.svg 14 KB
- assets/fundamentals-layers.svg 13 KB
- assets/hardening-frontier.svg 12 KB
- assets/hero.svg 6.0 KB
- assets/loop-costs.svg 10.0 KB
- assets/loop-thesis.svg 4.5 KB
- assets/loop-turn.svg 10 KB
- assets/loop.gif 3875 KB
- assets/red-team.svg 4.4 KB
- assets/router-cascade.svg 8.5 KB
- assets/router-pipeline.svg 7.4 KB
- assets/sdlc-loop.svg 3.6 KB
- assets/shepherd-loop.svg 12 KB
- assets/system-architecture.svg 8.1 KB
- scripts/apply-repos.sh 1.6 KB runs code
- scripts/check-actions.sh 6.1 KB runs code
- scripts/check-docs.sh 3.2 KB runs code
- scripts/check-mcp.sh 3.6 KB runs code
- scripts/check-vendor.sh 5.6 KB runs code
- scripts/check-workflows.sh 3.2 KB runs code
- scripts/compass-ablate.sh 10 KB runs code
- scripts/compass-audit-log.sh 2.0 KB runs code
- scripts/compass-audit-plugin.sh 17 KB runs code
- scripts/compass-bench.sh 11 KB runs code
- scripts/compass-dashboard.sh 6.9 KB runs code
- scripts/compass-digest.sh 9.2 KB runs code
- scripts/compass-drift.sh 4.1 KB runs code
- scripts/compass-enable.sh 13 KB runs code
- scripts/compass-gate.py 18 KB runs code
- scripts/compass-gate.sh 413 B runs code
- scripts/compass-impact.sh 3.9 KB runs code
- scripts/compass-listen.mjs 8.2 KB runs code
- scripts/compass-notify.sh 4.4 KB runs code
- scripts/compass-onboard.sh 9.2 KB runs code
- scripts/compass-policy-synth.sh 9.3 KB runs code
- scripts/compass-redteam-external.sh 17 KB runs code
- scripts/compass-redteam.sh 9.1 KB runs code
- scripts/compass-route.sh 9.3 KB runs code
- scripts/compass-sandbox.sh 2.7 KB runs code
- scripts/compass-sbom.sh 7.6 KB runs code
- scripts/compass-scan.sh 11 KB runs code
- scripts/compass-schedule.sh 21 KB runs code
- scripts/compass-spend.sh 5.4 KB runs code
- scripts/compass-status.sh 2.7 KB runs code
- scripts/compass-trace.sh 11 KB runs code
- scripts/compass-upgrade.sh 5.2 KB runs code
- scripts/content-corpus.tsv 5.3 KB
- scripts/context-pack.sh 4.5 KB runs code
- scripts/doctor.sh 7.8 KB runs code
- scripts/fixtures/audit-plugin/clean-plugin/.mcp.json 230 B
- scripts/fixtures/audit-plugin/clean-plugin/commands/format.md 152 B
- scripts/fixtures/audit-plugin/clean-plugin/hooks/format-on-edit.sh 395 B runs code
- scripts/fixtures/audit-plugin/clean-plugin/README.md 149 B
- scripts/fixtures/audit-plugin/compass-core.baseline 1.0 KB
- scripts/fixtures/audit-plugin/malicious-plugin/.mcp.json 403 B
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.
- 1mo ago First seen · 32 lines scan A 3a51c7a930e0
compass is a plugin published in the GitHub repository dshakes/compass (20 stars, last pushed 3d ago), licensed MIT. Its token cost is not measured: this kind of file is read by the harness, not the model. 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 2026-09-15.
Other plugins, from other repositories
catalysis-scientist
Reasons from active-site structure, turnover frequency, selectivity, and the Sabatier principle through CO/H2 chemisorption site-counting, Weisz-Prater and Mears transport checks, Langmuir-Hinshelwood/Mars-van Krevelen kinetics, and operando DRIFTS/XAS while treating diffusion-limited apparent rates, DRIFTS spectator…
decision-making
Master architectural decision-making: ADRs, technology selection, build-vs-buy analysis, proof of concepts, and migration planning.
claude-md-management
Tools to maintain and improve CLAUDE.md files - audit quality, capture session learnings, and keep project memory current.
humanizer
Rewrite AI-sounding text so it reads like the writer without changing what it says.
nextjs
Official Next.js skills: adopt and optimize Cache Components, adopt Partial Prefetching, and verify runtime behavior against a running dev server.
themes-example
themes-example. An open-source AI agent that brings the power of Gemini directly into your terminal.