research-loop

A command for running one step of an autonomous research loop, where a project repeatedly performs and records planned research tasks. It requires predictions, cost checks, verification, and written results before continuing.

In plain words
What is it for?
Use it to execute the next planned research step, record findings, verify the claimed outcome, update the research log, and schedule another iteration when appropriate.
Why use it?
It keeps long-running research work anchored to project instructions and makes positive, negative, and inconclusive results traceable.

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/research-loop
Clone the repo
git clone --depth 1 https://github.com/MissingPackage/nightshift
Per session 16 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 339 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 $0.00016 $0.00339
Opus 5 $0.00008 $0.00169
Sonnet 5 $0.00003 $0.00068
Haiku 4.5 $0.00002 $0.00034

Measured yesterday against content hash a3683cb3f0a3, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

research-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/research-loop.md · 20 lines

What it actually says

Continue the autonomous research loop for this project.

Follow the loop-iteration skill exactly. For this iteration:

  1. Re-anchor from disk: the PI directive files referenced in research/AGENDA.md (or HANDOFF.md), then AGENDA §1 "next decidable", then the PI docket.
  2. Execute the single next decidable research step. Standing constraints: pre-register the prediction in research/predictions.md BEFORE any --execute; dry-run and show cost estimates before real API spend; hard-stop and docket if projected spend exceeds the session budget in CLAUDE.md.
  3. Before scheduling the next iteration, spawn the loop-verifier agent with your claimed outcome. On FAIL: fix if decidable this iteration, otherwise docket and stop.
  4. Grade the outcome against the pre-registered prediction — honestly. Negative and null results are recorded with the same care as positives ("failures are data").
  5. Update: journal (append), AGENDA §1 (refresh in place), docket (append PI-gated items).
  6. Send the digest: 3-6 lines — what was done, why, what the results say, what's next.
  7. Schedule the next iteration (self-paced) — or STOP BY DESIGN if all remaining work is PI-gated, stating the docket items that need rulings. Never invent scope to keep the loop alive.

State machine and escalation semantics: ORCHESTRATION.md §4; goal spine (multi-phase research goals): §1-2.

$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 · 20 lines · 16 tokens per session scan A a3683cb3f0a3

Subscribe to this mod's changes

research-loop is a command published in the GitHub repository MissingPackage/nightshift (2 stars, last pushed 15d ago), licensed MIT. It adds 16 tokens to every session and 339 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.