product-loop

A guarded command for running one small product-development cycle at a time. It limits the work to one feature or fix, then requires implementation and checks before stopping.

In plain words
What is it for?
Use it to carry out one planned feature or bug fix on a separate branch, run tests and code-quality checks, and leave the result ready for review.
Why use it?
It prevents an automated coding agent from inventing large roadmap decisions, changing protected systems, or merging unfinished work. It also makes the completion checks explicit.

Command

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 commands/missingpackage/nightshift/product-loop
Clone the repo
git clone --depth 1 https://github.com/MissingPackage/nightshift
Per session 18 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 482 The whole file, excluding the scripts and references it only reads on demand.
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.

ModelPer sessionOnce invoked
Fable 5 $0.00018 $0.00482
Opus 5 $0.00009 $0.00241
Sonnet 5 $0.00004 $0.00096
Haiku 4.5 $0.00002 $0.00048

Measured yesterday against content hash fcd944defc35, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

product-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.

commands/product-loop.md · 21 lines

What it actually says

Continue the autonomous product loop for this project. Follow the loop-iteration skill exactly.

Standing guardrails (non-negotiable for every iteration):

  • Scope cap: ONE feature-slice or fix per iteration, small enough to implement AND verify (skill: done) within the iteration. Roadmap-scale choices about WHAT to build next come from HANDOFF.md §next-decidable / the tracker (Linear) — if neither names a next item, STOP and docket "roadmap ruling needed" instead of inventing features.
  • Protected paths: never touch deploy/infra manifests, auth/consent flows, or migrations unless the iteration's tracker item explicitly says so. Anything under the cofounder's ownership (deploy repo, cluster apps) is read-only: docket, don't edit.
  • Branch & merge policy: work on a feature branch per slice. NO merges to dev/main and NO pushes beyond the branch unless the goal text explicitly grants it. PR-ready is the target state; merging is a docket item by default.
  • Verification: full gates (skill: done) each iteration — unit + integration (real Docker) + lint + conventions (domain events, factory pattern, docs procedure). Run the consistency-sweep agent whenever the slice introduced or extended a pattern; INCOMPLETE coverage blocks the slice from PR-ready.
  • Residue routing: bugs/ideas found along the way → Linear/docket with one line each; never absorbed into the current slice.

State machine, escalation levels, and the unattended test: ORCHESTRATION.md §4. Goal spine: .harness/goals/<slug>/ (§1).

Each iteration: re-anchor (HANDOFF, tracker, docket) → pick the one named next slice → implement → verify (done-report saved into the journal) → loop-verifier gate → refresh HANDOFF §next-decidable → digest to the user (3-6 lines: slice, evidence one-liner, next) → schedule or stop-by-design.

Morning report: the final digest of a night run additionally lists: slices completed (with branches/PRs), the full docket delta, and anything a human must do before the next run.

$ARGUMENTS

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.

  1. yesterday First seen · 21 lines · 18 tokens per session scan A fcd944defc35

Subscribe to this mod's changes

product-loop is a command published in the GitHub repository MissingPackage/nightshift (2 stars, last pushed 15d ago), licensed MIT. It adds 18 tokens to every session and 482 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.