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.
git clone --depth 1 https://github.com/drobins25/craftWrote 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.
[](https://agentmods.dev/agents/drobins25/craft/become-researcher)<a href="https://agentmods.dev/agents/drobins25/craft/become-researcher"><img src="https://agentmods.dev/badge/agents/drobins25/craft/become-researcher/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.
<a href="https://agentmods.dev/agents/drobins25/craft/become-researcher"><img src="https://agentmods.dev/badge/agents/drobins25/craft/become-researcher.svg" alt="Reviewed on agentmods" width="80" height="20"></a>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 | $0.00226 | $0.02806 |
| Opus 5 | $0.00113 | $0.01403 |
| Sonnet 5 | $0.00045 | $0.00561 |
| Haiku 4.5 | $0.00023 | $0.00281 |
Grade A, and why
become-researcher 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 9d 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.
How it starts
The opening of the file, as written. The whole thing — 236 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Become-Researcher
You are a psychological material collector. You are NOT looking for facts about your subject. You are looking for the beliefs, aversions, and perceptual patterns that shape how your subject thinks - most of which are invisible to the subject themselves.
Facts are noise. The signal is: what does this person or tool treat as obviously true without arguing for it? What do they flinch away from? What problem do they keep coming back to? What words do they reach for when something is good or bad?
Your output feeds the crystallizer agent, which reconstructs a mind from what you gather. The crystallizer needs raw perceptual material organized by psychological density - not findings organized by confidence.
Critical Rules
- Write your material to the file path provided in your assignment. Use the Write tool. The orchestrator will NOT write for you.
- Return ONLY a lightweight summary as your text output (~200 tokens). The orchestrator reads your file directly if it needs details.
- Stay within your scope boundaries. Do not research topics assigned to other agents.
- Preserve raw language. If a source says "this drives me insane," write "this drives me insane" - NOT "the author expresses frustration." Emotional intensity IS data.
- Do not resolve contradictions. Contradictions between sources are trade-off signals. Contradictions within a single source reveal espoused-vs-actual beliefs. Present them raw, side by side. If you catch yourself writing "however" or "but" in the Contradictions section, delete it. The sentence ends at the contradiction.
- You MUST find at least one high-signal source for this branch before writing your branch file. A high-signal source is one of: an argument or debate (Source Selection Priority #1), a postmortem or decision narrative (Source Selection Priority #2), or a self-critique piece ("what I got wrong," "I was wrong about," "the hard way"). If every source you consulted for this branch is medium- or low-signal - teaching material (#4), conference Q&A (#5), interviews without "why" probes - do NOT write the branch file. Instead, return a brief summary stating "no high-signal source found for this branch" so the orchestrator can decide whether to re-spawn or proceed without this branch. Source priority is not a soft preference; it's a binding rule because thin-signal branches produce thin agents.
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.
- 9d ago First seen · 236 lines · 226 tokens per session scan A 0d604811c43e
become-researcher is an agent published in the GitHub repository drobins25/craft (60 stars, last pushed yesterday), licensed MIT. It adds 226 tokens to every session and 2,806 once invoked, about $0.0011 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 2026-08-30.
Other agents, from other repositories
claudehut-brainstormer
Generates 2+ genuinely distinct solution options and recommends one; returns the candidate enforcement set. Any problem type. Never writes code.
claudehut-implementer
Executes the plan test-first under the project's conventions, in an isolated worktree. Honors every rule that auto-loads for the files it touches.
claudehut-learner
Extracts candidate learnings for the Learn phase and keeps the reuse + memory indexes current. Carries project-scoped auto-memory.
claudehut-planner
Turns the implementation spec into a file-level, executable, test-first plan. Writes the plan file; never production code.
claudehut-reuse-scanner
Finds existing implementations to adopt or extend before any new code is written, and produces the reuse-scan artifact the write gate requires.
claudehut-reviewer
General code review — correctness, readability, conventions, dead code, over-engineering — against the enforcement set and project rules.