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.
/plugin marketplace add inferencegod/autonomy-loopnpx agentmods add plugins/inferencegod/autonomy-loop/marketplacegit clone --depth 1 https://github.com/inferencegod/autonomy-loopGrade A, and why
autonomy-loop 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 yesterday.
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
{
"$schema": "https://anthropic.com/claude-code/marketplace.schema.json",
"name": "autonomy-loop",
"owner": { "name": "inferencegod" },
"plugins": [
{ "name": "autonomy-loop", "source": "./autonomy-loop", "description": "Builder / reviewer / planner / researcher Claude Code terminals pass a git baton through a deterministic, frozen-invariant gate with an adversarial 5-lens reviewer and a mechanized bite. Presence-driven roles: launch a terminal to add one. A rings + provision-or-refuse security posture (a SessionStart preflight refuses unattended auto-promotion below a hardened trust tier, with a read-only control plane, a sandboxed shell, a repo-scoped credential, and a no-bypass prod ruleset), plus a greenfield mutation-bite, a money-path floor, and a coverage ratchet. Rigor on by default. Hardened across a 181-cycle self-dogfood through v0.8.4." }
]
}
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.
- yesterday First seen · 9 lines scan A 2e09cd7fb9ac
autonomy-loop is a plugin published in the GitHub repository inferencegod/autonomy-loop (1 stars, last pushed 2mo 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 2026-08-31.
Other plugins, from other repositories
merismos
Apportion an autonomous goal into coarse execution units and derive each unit's completion conditions — /apportion (μερισμός: an apportionment).
forge-harness
Plugin marketplace listing 2 plugins: fh-meta, fh-commons.
goalkeeper
Durable goal-following for Claude Code: contracts with definition-of-done, subagent judge gates, and linear chains.
polisade-marketplace
Plugin marketplace listing 1 plugin: polisade.
milestone-driver
Drives a GitHub milestone to completion: triages each issue for design gaps + dependency order up front, then for each buildable issue finds the root cause (or parks it), dispatches a TDD implementer subagent, runs unit + E2E tests + code review, opens a PR, and auto-merges issues on CI green. The whole loop sits…
shiploop
Single-plugin marketplace for shiploop — a self-improving multi-agent harness (every resolved ticket writes a lesson into your git-tracked CLAUDE.md) that grinds your ticket backlog across every repo in your product.