research

An explicit /research command for running measurable experiments through THINK, TEST, and REFLECT stages. It stores hypotheses, raw runs, reflections, and conclusions in a .lab/ directory.

In plain words
What is it for?
Use it for benchmarks, accuracy tests, latency comparisons, and other questions with a measurable success criterion. It repeats experiments when the results are too uncertain.
Why use it?
It provides a repeatable record for testing claims instead of reporting unsupported results, including sample size and variation when measurements are involved.

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/vrennat/developersdevelopers/research
Clone the repo
git clone --depth 1 https://github.com/vrennat/developersDevelopers
Per session 52 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 392 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 $0.00052 $0.00392
Opus 5 $0.00026 $0.00196
Sonnet 5 $0.00010 $0.00078
Haiku 4.5 $0.00005 $0.00039

Measured yesterday against content hash d64e2d33c849, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

research 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.md · 32 lines

What it actually says

/research

Autonomous experimentation for tasks with measurable outcomes. Use when you can quantify success: a benchmark, an accuracy score, a latency number, an A/B comparison.

Procedure

  1. Restate the question and the success metric in one sentence each.
  2. Create .lab/<slug>/ if it doesn't exist. This is the experiment workspace.
  3. THINK: form a hypothesis. Write it to .lab/<slug>/hypothesis.md with: claim, predicted measurement, smallest experiment that would prove or disprove it.
  4. TEST: run the experiment. Capture inputs, outputs, and the raw measurement to .lab/<slug>/runs/<timestamp>.md.
  5. REFLECT: compare measurement to prediction. Was the hypothesis confirmed, refuted, or unclear (variance too high)? Write to .lab/<slug>/reflections.md.
  6. If unclear: run more samples until variance is acceptable. Report sample size and confidence.
  7. Iterate from THINK with a refined or replacement hypothesis until the question is answered.
  8. Final report: .lab/<slug>/conclusion.md with the answer, the evidence, and the variance.

Rules

  • Never report a benchmark without sample size and variance.
  • If the result could plausibly be noise, say so. Don't rerun fishing for a better number.
  • Keep raw run data on disk; don't delete it.

When NOT to use

  • The answer is already known and just needs implementation: skip to /impl.
  • The question is qualitative (taste, style): brainstorming, not research.
  • Pure reading research: just read.
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 · 32 lines · 52 tokens per session scan A d64e2d33c849

Subscribe to this mod's changes

research is a command published in the GitHub repository vrennat/developersDevelopers (2 stars, last pushed 1mo ago), licensed MIT. It adds 52 tokens to every session and 392 once invoked, about $0.0003 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.