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/ethanaubuchon/dossier-tradecraftWrote 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/commands/ethanaubuchon/dossier-tradecraft/research)<a href="https://agentmods.dev/commands/ethanaubuchon/dossier-tradecraft/research"><img src="https://agentmods.dev/badge/commands/ethanaubuchon/dossier-tradecraft/research/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/commands/ethanaubuchon/dossier-tradecraft/research"><img src="https://agentmods.dev/badge/commands/ethanaubuchon/dossier-tradecraft/research.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 | $0.00023 | $0.01249 |
| Opus 5.5 | $0.00009 | $0.00500 |
| Sonnet 5.5 | $0.00005 | $0.00250 |
| Haiku 4.5 | $0.00002 | $0.00125 |
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 1mo 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 — 62 lines — stays where its author put it; the contents beside it link to each section on GitHub.
You are starting a research session. The topic (if provided): $ARGUMENTS
This is the /research recipe (Phase 1 v2) of the dossier-tradecraft framework.
Principles
- Vault = grounding, not a research target. The vault sweep enriches context for the web step; it is not presented alongside web findings as a parallel source.
- Web = primary research target, always runs. Model priors can't be trusted for fast-moving or landscape topics.
- Frame step is near-zero. Topic-as-stated passes through. Clarifications are conditional, not mandatory.
- Capture is agent-judged; write directly. You pick the slug, draft the note with citations, and write it. Surface what was captured afterward — the user can redirect, revise, or ask you to move/edit the note post-hoc. Anchor is a capture-time output, not a frame-time input.
- Loop iteration is context-accretive. Each turn builds on previous turn's captures.
- Thin caller. Cross-verify mechanics — cohort discipline, source-type annotations, used / marketplace consent gates, low retail-sample escalation — live in
agents/web-researcher.mdandskills/dispatch-exploration/SKILL.md. This recipe never re-implements them. Recipe-specific responsibilities are intake, loop shape, and capture; everything else is delegated.
Workflow
-
Load context — invoke the
load-contextskill (vault grounding pre-step). -
Intake. If
$ARGUMENTSis empty, ask what the user wants to research. Otherwise acknowledge the topic briefly and proceed. Do not interrogate scope, anchor, or biases up front. -
Pre-vault clarify (conditional). Skip unless the topic is genuinely ambiguous. Fire only when:
- A term has multiple common meanings and the right one isn't obvious from context
- The user mentions known candidates without specifying skim-worthy vs. deep-dive
- Scope is truly unclear (landscape scan vs. narrow comparison vs. how-to-build)
Default: skip. When in doubt, don't ask.
-
Vault grounding. Invoke
dispatch-explorationwith 1–3 vault queries (target: vault) scoped to what's likely to enrich the web step — the user's rig, prior thinking, related projects. Present the compressed findings briefly; frame them as context for the web search, not the research output.
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.
- 1mo ago First seen · 62 lines · 23 tokens per session scan A ec0711d4c18a
research is a command published in the GitHub repository ethanaubuchon/dossier-tradecraft (3 stars, last pushed 8d ago), licensed MIT. It adds 23 tokens to every session and 1,249 once invoked, about $0.0001 per session on Opus 5.5. 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 commands, from other repositories
token-cost-tracker
Estimate, log, and report token usage using runtime-supplied billing rates; never rely on embedded model prices.
wiki-lint
Run a health check on the wiki. Invoke with /wiki-lint or "lint the wiki".
daily-okr
Run a daily knowledge compound loop (7 KR). Invoke with /daily-okr or "start my daily review".
session-learn
Extract knowledge from the current session. Invoke with /session-learn or "extract this session".
help
Show everything the Obsidian Knowledge Agent can do, with every command and an example.
setup
Set up this folder as a knowledge vault — adapts to an empty folder or an existing vault.