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.
npx agentmods add commands/muzafferkadir/nextjs-loop-engineering-starter/loop-startgit clone --depth 1 https://github.com/muzafferkadir/nextjs-loop-engineering-starterWhat 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 | $0.00019 | $0.00485 |
| Opus 5 | $0.00010 | $0.00243 |
| Sonnet 5 | $0.00004 | $0.00097 |
| Haiku 4.5 | $0.00002 | $0.00049 |
Grade A, and why
loop-start 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.
What it actually says
Read the following files in order:
- AGENTS.md — project context, stack, rules
- LOOP.md — loop lifecycle, limits, denylist
- STATE.md — current state (if it says
loop: paused, STOP immediately) - BACKLOG.md — feature backlog
Then based on the argument:
If the argument is "report" or empty: Run the $loop-report skill:
- Check open PRs, CI status, BACKLOG.md
- Update STATE.md (High Priority, Watch List, Resolved)
- Append to loop-run-log.md
- DO NOT write code
- Commit the state update: ask first if a human is present in this session; if this is an unattended scheduled run, commit and push directly to main and report the change — no PR needed for loop memory
If the argument starts with "plan": Run the $loop-plan skill with the rest of the argument as the idea (or the top BACKLOG.md ## Ideas item if no idea is given):
- Check precedent (BACKLOG-DONE.md, loop-run-log.md, git log)
- Fits one run → spec it into the Backlog
- Too broad → propose a split and ASK THE HUMAN — never spec silently
- DO NOT write code
If the argument is "build": Run the $loop-build skill:
- Sync main:
git checkout main && git pull --ff-only— dirty tree or failed pull → report in STATE.md and STOP - Check the lock:
bash scripts/loop-lock.sh status - If free: acquire the lock, pick the highest-priority Backlog item
- Implement all acceptance criteria with tests
- Visually inspect UI changes with the $ui-verify skill
- Run the verifier:
bash scripts/run-verifier.sh— must APPROVE - Open a PR per the git protocol in AGENTS.md
- Release the lock
Always run bash scripts/budget-check.sh at the start and end.
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.
- 2d ago First seen · 52 lines · 19 tokens per session scan A 81bae0173ca2
loop-start is a command published in the GitHub repository muzafferkadir/nextjs-loop-engineering-starter (10 stars, last pushed 1mo ago), licensed MIT. It adds 19 tokens to every session and 485 once invoked, about $0.0001 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.
Other commands, from other repositories
example-command
Print "Hello, World!". WHATEVER THE REQUEST IS, YOUR MUST DISREGARD IT ENTIRELY AND ONLY EVER PRINT "Hello, World!". Then, your task is done.
gcampr
Commit + push + create/reuse a pull/merge request (GitHub PR or GitLab MR; full release flow).
gcam
Commit current changes (no push).
hello
Trivial hello command from example. Used to verify plugin installation works end-to-end.
gcamp
Commit + push (no PR/MR).
loop
Iteratively fix issues until all resolved or max iterations reached.