shiphrahx/ai-for-engineering-leaders/difficult-conversation-prep
Claude Code and agent skill by shiphrahx/ai-for-engineering-leaders
- community
Checking that the SKILL.md file is still in the repository.
Produces talking points for a hard conversation with a direct report — an opening frame, a direct evidence-based core message (no sandwich method), an impact statement, space to listen, a path forward with specific expectations, a support offer, and a relationship-affirming close. Use when the user says "help me prep a hard conversation", "I need to give tough feedback", "address a conflict", or describes a behavioural concern. Use this to plan delivery of feedback; use pip-documentation when i…
Install
npx vskill@latest install shiphrahx/ai-for-engineering-leaders/difficult-conversation-prepSKILL.md
Difficult Conversation Prep
Structure talking points so the conversation is clear, kind, and specific — the person leaves understanding exactly what the issue is, why it matters, and what "better" looks like, without feeling attacked.
Inputs to gather
Gather these before drafting. If any are missing, ask in a single batched question — never invent the examples, the evidence, or what was said. Mark anything genuinely unavailable as Unknown.
- Person — name and role
- What the conversation is about — the specific issue
- Specific examples — dated, observable evidence
- Desired outcome — what you want to change
- Relationship context — your relationship and how they respond to feedback
Steps
- Write an Opening Frame — how to start: set the tone and state the purpose. Frame it as "your impact isn't matching your intent", not "you're doing something wrong".
- Write the Core Message — the specific issue stated directly with dated, observable examples. No sandwich method, no interpretations ("you're rude") — just what happened and the effect.
- Write the Impact Statement — why it matters to the team, the project, and their own career.
- Write a Listen prompt — genuine, open questions inviting their perspective, with a note to actually pause and hear them out before moving to solutions; they may have context you don't.
- Write the Path Forward — specific, observable expectations for what changes.
- Write the Support Offer — concrete ways you'll help (calibration, a model peer to learn from, pairing).
- Write the Close — reaffirm the relationship and check how they're feeling; this comes from respect.
- Adapt to context: if the person gets defensive, don't argue — restate the facts and that you're asking for an adjustment, not agreement with the interpretation. For a missed-promotion conversation, lead with empathy, be specific about what was missing, and pivot to a clear path forward. For a conflict between two reports, talk to each individually first; don't open with a joint mediation. For the manager's own nervousness, rehearse the opening sentence out loud. Always follow up in writing within 24 hours ("here's what we agreed to") to create accountability.
- Assemble the output in the format below, with brief parenthetical coaching notes on delivery.
Output format
Boundaries
- Never invent examples, evidence, quotes, or what was said. Mark unknowns as Unknown and ask — vague feedback without specifics is the opposite of the goal.
- State observable behaviour and impact, never character judgments ("you're lazy/rude/difficult"). Keep it about what happened and what changes.
- Base the conversation on documented evidence, not impressions; ensure the standard is one applied to peers, free of bias.
- For sensitive or potentially employment-affecting topics (a behavioural pattern heading toward a PIP, or anything legal), involve HR — and never use this to deliver a predetermined decision dressed as a discussion.
- Keep prep notes confidential; have the conversation in person or video, never by message, and never repeat one report's words to another.
Chaining
- If the issue is a documented performance pattern that coaching hasn't fixed, offer pip-documentation (with HR).
- After the conversation, capture the agreed path forward in the next performance-review-draft or follow up in one-on-one-prep.
Security
- Security scanPassed, 100 of 100Pattern checks on the skill's files.
- Model reviewNot runA model review is scheduled for the next rescan.
Quality evaluation
- Task evaluationNot runNo evaluation yet. Run one in Skill Studio.
Source
- SKILL.md in the repositoryCheckingskills/people-management/difficult-conversation-prep/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/shiphrahx/ai-for-engineering-leaders/difficult-conversation-prep)