What is supervised-loop?

A bounded automation loop for making one small, supervised change in the Guitar Alchemist repository, then running its checks and recording evidence.

Use it to implement and validate a single eligible improvement in the repository's permitted areas.

It limits autonomous work to one approved roadmap slice and requires preflight checks and verification before stopping.

supervised-loop — Skill for Claude Code by GuitarAlchemist/ga

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.

Any agent

npx skills add GuitarAlchemist/ga --skill supervised-loop

Clone the repo

git clone --depth 1 https://github.com/GuitarAlchemist/ga

Made for: Claude Code.

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/skills/guitaralchemist/ga/supervised-loop/github.svg)](https://agentmods.dev/skills/guitaralchemist/ga/supervised-loop)

Your own site

<a href="https://agentmods.dev/skills/guitaralchemist/ga/supervised-loop"><img src="https://agentmods.dev/badge/skills/guitaralchemist/ga/supervised-loop/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.

Your own site · 80×15

<a href="https://agentmods.dev/skills/guitaralchemist/ga/supervised-loop"><img src="https://agentmods.dev/badge/skills/guitaralchemist/ga/supervised-loop.svg" alt="Reviewed on agentmods" width="80" height="20"></a>

Per session 85 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.

When invoked 2,552 The whole file, excluding the scripts and references it only reads on demand.

Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe.

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 5.1	$0.00085	$0.02552
Opus 5.5	$0.00034	$0.01021
Sonnet 5.5	$0.00017	$0.00510
Haiku 4.5	$0.00009	$0.00255

Measured 1mo ago against content hash 3beba1b64d64, method: parsed. Prices are Anthropic first-party input rates as of 2026-10-07, from the pricing page.

Security

Grade A, and why

supervised-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 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.

.claude/skills/supervised-loop/SKILL.md · 235 lines

How it starts

The opening of the file, as written. The whole thing — 235 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Supervised Autonomous Loop (GA)

Operationalises the supervised-loop pattern from agent-blackbox on the Guitar Alchemist repo. The cycle is intentionally short: one slice, one verify, one evidence file, then stop.

GA-specific scope

GA has a 5-layer strict bottom-up architecture (per CLAUDE.md):

Core — GA.Core, GA.Domain.Core
Domain — GA.Business.Core, GA.Business.Config, GA.BSP.Core
Analysis — GA.Business.Core.Harmony, GA.Business.Core.Fretboard
AI/ML — GA.Business.ML (embeddings, OPTIC-K, RAG)
Orchestration — GA.Business.Core.Orchestration, GA.Business.Assets, GA.Business.Intelligence, Apps

The supervised loop only operates within layer 4/5 + frontend/docs/scripts on the initial rollout. Layers 1-3 (Common/GA.Business.Core/** and below) are in protected_paths because a regression there fans out across every consumer. The OPTIC-K embedding dimension (240-dim) is a one-way door and is also protected. Widen the scope only after several clean cycles and explicit operator review.

When to use

The user asked for one bounded autonomous improvement on this repo.
You are running a scheduled /loop and need to make one safe slice of progress before yielding.
You are dogfooding the supervised pattern that other repos will copy.

Hard refusals (never bypass)

state/governance/dev-process-overseer.json missing, stale (>24h), or workflowMode != "loop-eligible" -> abort with reason.
.STOP file at repo root -> abort.
HALT-ALL marker present at $HOME/.demerzel/HALT-ALL and active -> abort.
ga.loop-policy.json missing allow_edit or protected_paths -> abort. (The risk-scoring policy lives in agent-blackbox.policy.json; the loop scope lives separately to keep the risk-scoring policy pristine.)
Any diff touches a path in protected_paths -> abort, do not commit.
pwsh Scripts/supervised-loop-preflight.ps1 emits LOOP_READY=false -> abort with the reason it printed; do not try to repair the gate. The preflight enforces six gates (loop-policy parses, overseer fresh, overseer loop-eligible, no .STOP, no HALT-ALL, worktree clean). The worktree-clean gate forgives dirty files outside both allow_edit and protected_paths but still hard-fails on dirty files INSIDE either.
Service restarts (Restart-Service, taskkill against GaApi / GaChatbot.Api / cloudflared / Ollama / Redis), secret rotation, cross-repo schema edits (docs/contracts/**, ix schemas, Demerzel governance schemas), OPTIC-K dimension changes -> abort, hand back to human.

Read the full file on GitHub · 235 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.

1mo ago First seen · 235 lines · 85 tokens per session scan A 3beba1b64d64

Subscribe to this mod's changes

supervised-loop is a skill published in the GitHub repository GuitarAlchemist/ga (2 stars, last pushed yesterday), licensed MIT. It adds 85 tokens to every session and 2,552 once invoked, about $0.0003 per session on Opus 5.5. 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 2026-09-03.

Project management Testing

team-qa

Orchestrate the QA team through a full testing cycle — qa-lead strategy and test plan, qa-tester case writing, execution, sign-off.

Donchitos/Claude-Code-Game-Studios · 35 tokens

scaffold-dotnet-test-project

MUST USE when an existing .NET test project was excluded from a .slnf/CI solution filter, disappeared from .sln/.slnx discovery, or lost its production ProjectReference; also for requests to set up, create, reuse, add, register, include, or repair a test project. Handles "tests pass directly but CI discovers zero"…

dotnet/skills · 138 tokens

test-result-feedback

Close the acceptance loop for a tested change — consume the apc test TierResult, map EACH acceptance criterion to a pass/fail outcome from structured fields (never chat text), attach an evidence ref, and write it back to the task. A SUT red is a finding, never softened to green.

Prismer-AI/PrismerCloud · 65 tokens

supervised-loop