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.
npx skills add richer-richard/cc-research --skill research-methodgit clone --depth 1 https://github.com/richer-richard/cc-researchWrote 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/skills/richer-richard/cc-research/research-method)<a href="https://agentmods.dev/skills/richer-richard/cc-research/research-method"><img src="https://agentmods.dev/badge/skills/richer-richard/cc-research/research-method/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/skills/richer-richard/cc-research/research-method"><img src="https://agentmods.dev/badge/skills/richer-richard/cc-research/research-method.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 | $0.00116 | $0.04755 |
| Opus 5 | $0.00058 | $0.02377 |
| Sonnet 5 | $0.00023 | $0.00951 |
| Haiku 4.5 | $0.00012 | $0.00475 |
Grade A, and why
research-method 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 8d 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 — 124 lines — stays where its author put it; the contents beside it link to each section on GitHub.
cc-research — research method (the loop)
This skill defines the orchestration loop. You are the orchestrator. You plan the work, dispatch subagents via the Task tool, triangulate what they find, and have the synthesizer write a cited report. You persist all evidence to disk so the run survives compaction and can be resumed.
Narrate every phase in plain prose to the user as you go — narration is the primary way progress is visible in the terminal. Use short status lines like the ones quoted in each phase below.
Models (pinned)
- Orchestrator: you (the calling model).
researcher→ Sonnet ·local-researcher→ Haiku ·synthesizer→ Opus.
Subagents may appear namespaced as cc-research:researcher, cc-research:local-researcher, cc-research:synthesizer. Dispatch them with the Task tool by that role; if a bare name is not found, use the cc-research:-prefixed name.
Modes (selected by the first argument)
- full —
<query>: run the whole loop and producereport.md. - plan — first token is exactly
planthen a query: do Scope + Plan only (writemeta.json+plan.md), then STOP. Do not retrieve. Setexpect_report:false. - resume — first token is exactly
resumethen a run id: re-openruns/<id>/, readmeta.json+plan.md+ any existingsources-*.jsonl/claims-*.jsonl, and continue from the first unfinished phase. If you proceed past planning to produce a report (the normal case — e.g. resuming aplan-only run), setexpect_report:trueinmeta.jsonso the Cite audit applies to this run.
Reserved-word rule: plan / resume are special only when they are the exact first token followed by a space. Anything else (including a query that merely contains those words later) is treated as a full-loop query.
Budget rails (overridable; there is no default dollar cap)
- No dollar cap by default. Runs are uncapped so a deeper, more prolonged search can spend what it needs. A user may set an optional cap with
--budget <usd>(SDK) — only then is a hard limit enforced. - maxTurns: 60 (overridable with
--max-turns). - ≤ 5 web researchers per wave, ≤ 2 waves (overridable with
--max-waves). - ≤ 8 tool calls per researcher — a self-imposed budget each researcher follows from its prompt. Its hard backstop is the agent's
maxTurns: 16(honored by both the plugin and the SDK) — set well above the soft budget so it catches only a runaway, not a normal deep strand. It caps agentic turns, not tool calls (a single turn can batch parallel calls), so the ≤ 8-call budget is the everyday governor and 16 turns is the safety ceiling. - ≤ 3 local researchers total.
- "Aim wide" operates within these structural rails: dispatch as many independent strands as the query warrants up to the per-wave ceiling, then use a second wave for the rest.
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.
- 8d ago First seen · 124 lines · 116 tokens per session scan A ff385ee6fae6
research-method is a skill published in the GitHub repository richer-richard/cc-research (1 stars, last pushed 3mo ago), licensed MIT. It adds 116 tokens to every session and 4,755 once invoked, about $0.0006 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 skills, from other repositories
returns-policy
Answer return, refund, and warranty questions for electronics. Use when the user mentions returns, refunds, RMAs, warranty coverage, damaged items, or opened packaging.
technical-troubleshooting
Provide setup, troubleshooting, and maintenance guidance. Use when the user reports a device that won't power on, connectivity issues, setup questions, overheating, or maintenance concerns.
scout-briefing
Launch a Scout morning briefing session as a background process. Auto-detects mode (morning briefing, consolidation, or weekend briefing) from the current time and day.
scout-consolidation
Launch a Scout consolidation session as a background process. Runs a lighter delta scan — captures what changed since the last run and updates action items and KB.
scout-research
Launch a Scout research session as a background process. Goes outward — discovers new information about KB entities via web search, docs, and APIs, then integrates findings.
scout-dream
Launch a Scout dreaming session as a background process. Processes feedback, does KB deep work, and works on wishlist items.