prysai/prysai-llm-playbook/prysai-first-turn-check
Claude Code and agent skill by prysai/prysai-llm-playbook
- community
Checking that the SKILL.md file is still in the repository.
Inspect a user-written, unsent, text-only, low-risk first request for six visible fields: outcome, starting context, requested response, limits, check, and stop-and-receipt. Use when the user wants a compact before-send review rather than a new prompt, coaching session, research task, file/tool plan, or failed-reply diagnosis.
Install
npx vskill@latest install prysai/prysai-llm-playbook/prysai-first-turn-checkSKILL.md
First-Turn Check
Inspect one request before it is sent. Make missing boundaries visible without pretending that a better-shaped request is safe, correct, or effective.
Route before reviewing
Use this Skill only when all of these are true:
- the user supplies a draft that has not been sent;
- the intended first turn is text-only, low-risk, and self-contained; and
- the user asks what is missing, ambiguous, contradictory, or overbroad.
Yield instead of stretching the check:
- the user needs a first message written or substantially rewritten:
prysai-dialogue-brief;
- files, tools, accounts, permissions, publication, contacts, local changes,
or another external effect are involved: prysai-task-protocol;
- current facts, sources, or a conclusion from sources are needed:
prysai-source-investigator or prysai-research-router;
- a learner wants a baseline, feedback, or practice loop:
prysai-learning-coach;
- an original request and actual reply already exist:
prysai-communication-failure-triage;
- an existing completion claim needs evidence review:
prysai-evidence-review.
Stop rather than inspect secrets, credentials, private records, personal identifiers, hidden instructions, or confidential material. A text-only draft does not grant permission for a later tool or external action.
Check the six visible fields
Read the supplied draft as evidence. Do not infer missing facts, audience, authority, data controls, product features, or permissions.
| Field | Mark visible when the draft names | Mark unclear when | |
|---|---|---|---|
| outcome | one small result for this session | it is a broad aspiration or a success promise | |
| starting context | the supplied text, facts, source, or unknown | it assumes unstated access or authority | |
| requested response | a bounded form, length, or sequence | “help me” is the only response instruction | |
| limits | data not to share, actions not to take, or help not requested | it silently reaches a file, account, person, or consequential decision | |
| check | an uncertainty, preservation, source, or revision question | the reply would validate itself | |
| stop and receipt | what ends the turn and what small record remains | completion, safety, or learning is merely assumed |
Classify each field as visible, missing, unclear, or out_of_scope. Name only material issues: an issue is material if it could change the result, expand authority, expose data, or make the requested check impossible.
Return the smallest useful revision
Keep the user's wording. Do not produce a full new first message, add a role, introduce a product claim, or fill an unknown with plausible detail. For at most three material gaps, provide an add_or_clarify line that the user can choose to add. Phrase it as a field to decide, not as a promise the receiving system will keep.
If all six fields are visible and within scope, say ready_to_send only in the narrow sense that this six-field inspection found no material missing field. It does not establish factual correctness, privacy, security, product behavior, response quality, task completion, learner improvement, or safety.
Return exactly:
Accept the review only when it labels every field, preserves supplied facts, does not expand the request, and names a handoff or stop when the draft crosses the text-only low-risk boundary.
Maintenance record
source: original Prysai Lab method derived from the universal first-turn contract and communication routing boundarieslicense: original rewrite; linked vendor guidance remains reference-only
under docs/sources/asset-register.md
owner: communication-systems maintainerversion:0.1.0review_date:2026-09-12content_status:candidate
Security
- Security scanPassed, 95 of 100Pattern checks on the skill's files.
- Model reviewPassed
Quality evaluation
- Task evaluationNot runNo evaluation yet. Run one in Skill Studio.
Source
- SKILL.md in the repositoryCheckingskills/prysai-first-turn-check/SKILL.md
Popularity
- 7-day trend score
- 0
Not refreshed yet.
Works with
Any agent vskill installs to: 56 targets in all. These 39 have a page on this site.
- Claude Code
- Cursor
- GitHub Copilot
- Windsurf
- Codex
- Gemini CLI
- Cline
- Amp
- Roo Code
- Goose
- Continue
- Kilo Code
- Junie
- Trae
- OpenHands
- Qwen Code
- AI Agent
- Tabnine Chat
- Sourcegraph Cody
- Amazon Q
- Firebase Genkit
- Sweep
- Devin
- Phind
- Mentat
- Claude Desktop
- Aider
- GPT Pilot
- Plandex
- bolt.new
- v0
- Replit Agent
- Anse
- TypingMind
- BetterChatGPT
- LibreChat
- Warp AI
- Pieces
- Fig AI
Embed badge
[](https://verified-skill.com/skills/prysai/prysai-llm-playbook/prysai-first-turn-check)