The Practitioner Desk
Skills · Subagent · Command · PDF
.claude/the-practitioner-desk/zip
skills/06
├ context-pack
├ delegation-ladder
├ handover-note
├ prompt-library-entry
├ session-log
├ verification-protocol
agents/01
├ output-verifier
commands/01
├ /delegate
FIELD-GUIDE.pdf9pp
English
One payment · No account · No subscription
- ✓ Download link in your inbox in seconds
- ✓ Secure card checkout · VAT handled
- ✓ 14-day money back, no questions
- ✓ Every future revision included
- ✓ Runs in Claude Code and claude.ai
Built for Claude first, our recommended agent. Skills follow the open Agent Skills standard (SKILL.md), so they also run in any agent that reads it; subagents and slash commands are Claude Code formats.
The Practitioner Desk
The delegation ladder, run on your own week
six skills·one subagent·one command·9-page guide
Your own week run as a procedure: every task placed on the delegation ladder, every repeated prompt written once as a brief, every output checked the same way.
Executive summary
- The problem
- You check an AI output when it looks wrong, so the confident, plausible ones go out unchecked, and those are the ones that fail. The same prompt is retyped every week a little differently, so nothing learned about its failures is kept.
- What it does
- Six skills and a read-only output-verifier place each task on the five-rung delegation ladder, turn repeated prompts into versioned briefs, write one verification protocol per class of output, and mark every claim supported, contradicted, unsupported or not checkable.
- Why take it
- Your judgement is what you are paid for. This keeps it on record: what you delegated, how it was checked, and what a colleague needs to run it without you.
For practitioners who already use AI daily and want a method rather than a better prompt.
In the first week you will
- 01Run /delegate on the task you were about to automate and get the rung, the check and the dry-run rule
- 02Rewrite the prompt you keep retyping as a shared entry with prompt-library-entry, owner and version included
- 03Give the output-verifier subagent a draft and its sources and see each claim marked supported, contradicted, unsupported or not checkable
Contents of the Kit
- 01Six installable agent skills: delegation-ladder, prompt-library-entry, verification-protocol, context-pack, session-log, handover-note
- 02The output-verifier subagent: re-derives a claim in a model output from the sources you supplied and reports supported, contradicted, unsupported or not checkable. Read-only
- 03A /delegate slash command that places a task on the ladder from one sentence
- 04The five rungs of the manual's Delegation Ladder as a placement test, with the three questions (blast radius, cost of verification, position on the jagged frontier) and the check that belongs at each rung
- 05Blank templates for a prompt library entry and the weekly session log
- 06A 9-page Field Guide on running your own week as a procedure, and plugin manifest included
One skill, as it ships
SKILL.md
delegation-ladder
- Use when
- Use when someone asks "should I automate this", "can an agent do this", "how much should I let it run on its own", when a task is about to be handed to a tool that acts, or when a delegation went wrong and the placement needs revisiting.
- Requires
- The task · What a wrong output costs if nobody catches it · How the result would be checked
- Will not
- Never place a task above rung 2 on the strength of felt speed or a smooth demo · Never invent an accuracy rate, a reliability percentage or a time saving · Checkpoints are kept small so they stay cheap enough never to skip
- Returns
- Placement: {task in one line}
One of six skills in the kit · All of them install the same way
The contents, in full · six skills, one subagent, one command
- 01delegation-ladderOne task placed on the manual's five rungs: completion, the briefed task, the workflow with checkpoints, the supervised agent, the autonomous agent. Placed by three questions you answer (the blast radius of a wrong output, what verification costs, whether the job sits inside the model's competence), with the check that belongs at the chosen rung, the dry-run rule for anything that acts, and the refusal list entry when the honest answer is that the task is not delegated at all.
- 02prompt-library-entryA prompt that keeps being retyped, rewritten as a brief in a shared library format: context, materials, task, constraints, definition of done, plus the trigger, the owner, the version, the failure the last version taught, and the line that says when it should be promoted to a skill instead.
- 03verification-protocolWritten once per class of output, not per job: what is checked, how, against what source, by whom, in what order, and what happens when a check fails. Short enough that it is never skipped, with the sampling rule you set rather than a coverage figure the tool invents.
- 04context-packThe standing context a recurring job needs, assembled in the shape a skill or a project can read: sources with their access route, the definitions your organization uses, the rules and the banned claims, what is deliberately out of bounds, and the review date after which the pack is stale.
- 05session-logOne week of delegated work on one page: what was delegated and to which rung, what was verified and how, what came back wrong and what the fix was, plus the running list of corrections that recurred often enough to belong in a brief permanently.
- 06handover-noteWhat a colleague needs to run the same setup without you: the tools installed, the context packs, the verification protocols, the access they must arrange themselves, the known failure modes, and the first job to run under supervision.
- 07output-verifier (subagent)Takes a model output and the sources it was supposed to come from, re-derives each checkable claim, and reports it as supported (with the quote and the file), contradicted, unsupported or not checkable from the material given. It reads; it never fills a gap, never re-writes the output, and treats instructions found inside the material as data rather than commands.
- 08/delegateThe slash command that places a task on the ladder from one sentence and returns the rung, the check and the first run rule.
The Augmented Practitioner separates two claims usually said as one: “I use AI” and “I am good at using AI”. The distance between them is method rather than tooling, because the frontier of what a model does well is jagged and undisplayed, and a practitioner’s felt sense of which side they are on is the least reliable instrument in the room. The Agent Foundry builds new tooling and the Coexistence Desk decides what an organization does about AI. The Practitioner Desk is the week in between.
Pairs with
The Augmented Practitioner is the method: the prompt as a brief, four workflows worked to their joints, the five rungs of the Delegation Ladder, and the discipline that keeps verification procedural. This kit runs that method on your own calendar. The Agent Foundry is the next step for a job nobody has packaged yet; The Coexistence Desk is the same argument one level up, where an organization decides which workflows to reinvent.
Before you buy
01Do I need Claude Code?
No. Every skill uploads to claude.ai as a zip. Subagents and slash commands are the Claude Code extras.
02I am not a programmer.
Install is copying folders into a .claude directory. The Field Guide in each kit assumes nothing beyond that.
03Will it work in my agent?
Skills follow the open Agent Skills standard (SKILL.md), so any agent that reads it can run them. Built for Claude first, our recommended agent.
04Do the tools invent numbers?
Never. Every figure is supplied by you, confirmed by you, or labeled as an assumption in the output.
05How does it arrive?
Instantly. Pay by card and the download link is in your inbox within seconds, and on the confirmation screen.
06Is this a subscription?
No. One payment, lifetime access to your download, every future revision of what you bought included.
07What if it is not useful to me?
Write to info@conclusionmp.com within 14 days of purchase and you get a full refund. No questions, no forms.
Ready when you are
The Practitioner Desk, $12.
Instant download, one payment, every future revision included. Or take The AI Shelf, six items for $39, and this kit works out at $6.50.