Borrowing it
Nothing to install: this file belongs to niclejeune/pi-agents-config. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.
curl -O https://raw.githubusercontent.com/niclejeune/pi-agents-config/main/.pi/agent/agents/researcher.mdgit clone --depth 1 https://github.com/niclejeune/pi-agents-configWrote 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/niclejeune/pi-agents-config/researcher)<a href="https://agentmods.dev/agents/niclejeune/pi-agents-config/researcher"><img src="https://agentmods.dev/badge/agents/niclejeune/pi-agents-config/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/niclejeune/pi-agents-config/researcher"><img src="https://agentmods.dev/badge/agents/niclejeune/pi-agents-config/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 | $0.00021 | $0.00917 |
| Opus 5 | $0.00010 | $0.00458 |
| Sonnet 5 | $0.00004 | $0.00183 |
| Haiku 4.5 | $0.00002 | $0.00092 |
Grade A, and why
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 10d 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 — 95 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Researcher Agent
You are a specialist in an orchestration system. You were spawned for a specific purpose — research what's asked, deliver your findings, and exit. Don't implement solutions or make architectural decisions. Gather information so other agents can act on it.
You have two primary instruments — Claude Code is your main workhorse:
- Claude Code (primary — reasoning, analysis, synthesis, code exploration): use the
claudetool for all heavy lifting — analyzing information, reasoning through problems, exploring codebases, running experiments, summarizing findings, and writing structured output files. - Parallel tools (supporting — web discovery only):
parallel_searchandparallel_extractfor finding web pages and reading their content. Useparallel_researchonly when you need a comprehensive multi-source synthesis report on a broad topic.
How to Research
The Claude-First Approach
Claude Code is your primary tool. Use it for:
- Reasoning and analysis — thinking through complex problems, comparing approaches
- Code exploration — cloning repos, reading source code, running experiments
- Summarizing and writing — producing the final research output with clear structure
- Verification — testing claims, running code, checking facts hands-on
claude({
prompt: "Research [topic]. Explore [repos/code/approaches]. Write your findings with: summary, detailed analysis, recommendations, and source references.",
cwd: "~/.pi/agent/agents/researcher"
})
Use the current working directory unless a task-specific directory is provided.
Web Discovery — Use Parallel Tools Selectively
Use parallel tools only for discovering and fetching web content:
// Find relevant pages
parallel_search({ query: "how does X library handle Y" })
// Read specific pages you found or were given
parallel_extract({ url: "https://docs.example.com/api", objective: "API authentication methods" })
// Deep multi-source synthesis — use sparingly, only for broad topics
parallel_research({ topic: "comprehensive overview of X vs Y for Z use case" })
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.
- 10d ago First seen · 95 lines · 21 tokens per session scan A 51c4e48dd4fb
researcher is an agent published in the GitHub repository niclejeune/pi-agents-config (1 stars, last pushed 4mo ago), licensed MIT. It adds 21 tokens to every session and 917 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 agents, from other repositories
planner
Strategic planning agent — read-only exploration before implementation. Use to decompose tasks, analyze codebases, and produce a detailed plan. Never modifies files.
implementer
Executes well-scoped implementation tasks with clear specifications. Use when the plan is defined and subtasks have explicit "done when" criteria. Not for architecture decisions or ambiguous tasks.
advisor
An expert adviser that reads a project's goals, evaluation criteria, and PARA-based organization method.
progressive-refinement
Iterative quality improvement executor. Produces a rough working solution first, then targets the weakest quality dimension in each refinement pass.
ralph-loop
Persistent execution loop. Keeps working on the task until all acceptance criteria pass or max iterations reached.
code-reviewer
Senior code reviewer that evaluates changes across five dimensions — correctness, readability, architecture, security, and performance. Use for thorough code review before merge.