AgentStack
SKILL verified MIT Self-run

Difficult Conversations

skill-skyf0xx-better-thinking-difficult-conversations · by skyf0xx

>

No reviews yet
0 installs
19 views
0.0% view→install

Install

$ agentstack add skill-skyf0xx-better-thinking-difficult-conversations

✓ scanned · ✓ verified — works with Claude Code, Cursor, and more.

Security review

✓ Passed

No issues found. Passed automated security review. · v0.1.0 How review works →

  • Prompt-injection patterns
  • Secret / credential exfiltration
  • Dangerous shell & filesystem operations
  • Untrusted network calls
  • Known-malicious package signatures

What it can access

  • Network access No
  • Filesystem access No
  • Shell / process execution No
  • Environment & secrets No
  • Dynamic code execution No

From automated source analysis of v0.1.0. “Used” means the capability is present in the source — more access means more to trust, not that it’s unsafe.

Are you the author of Difficult Conversations? Claim this listing to set pricing, connect Stripe payouts, and keep 70% of every sale.
Sign up to claim

About

Difficult Conversations

Prepare for and conduct a high-stakes conversation — sharing a hard truth, addressing a conflict, delivering unwelcome news — without sacrificing the relationship or dodging the substance.

Why

Difficult conversations fail in two opposite directions: avoided entirely, so the issue festers, or delivered as an ambush, so the relationship breaks. Preparation that separates the facts from the story you've told yourself, plus a structure that invites the other side in, threads that needle.

Use when / Don't use when

  • Use when: a conversation has been avoided because it's uncomfortable but the cost of continued avoidance is rising; conflict has an emotional charge that a purely factual message would mishandle.
  • Don't use when: the matter is simple enough for direct, low-stakes feedback — feedback-delivery alone suffices.

Inputs → Outputs

  • Inputs: a conversation both necessary and likely to be emotionally charged.
  • Outputs: a conducted conversation that addressed the substance and preserved, or appropriately recalibrated, the relationship.

Principles

  • Separate the facts of what happened from the story you've constructed about why — the story is usually where the conversation goes wrong.
  • Every difficult conversation carries three layers: the "what happened" content, the feelings involved, and each party's sense of identity at stake.
  • The goal is usually mutual understanding and a path forward, not "winning" the account of events.

Procedure

  1. Prepare: separate observable facts from your interpretation or story about them. Run empathy-mapping on the other party — what do they likely see, feel, need?
  2. Run steelmanning on their probable position — enter the conversation able to state it fairly.
  3. Identify your actual goal — understanding, a changed behavior, a repaired relationship, a decision — be specific.
  4. Open: state the issue using nonviolent-communication structure — observation, feeling, need — avoiding the story or interpretation in the opening.
  5. Listen: use active-listening genuinely — invite and absorb their account before defending your own.
  6. Navigate identity: if either party's sense of competence, worth, or belonging feels threatened, name that directly rather than letting it drive the conversation from underneath.
  7. Close: summarize the shared understanding, even if disagreement remains, and agree on a concrete next step or request.

Common mistakes

  • Opening with the story or interpretation instead of the observable facts, triggering immediate defensiveness.
  • Preparing your talking points but not genuinely preparing to listen.
  • Avoiding the identity layer, so it silently sabotages an otherwise well-structured conversation.

Examples

  • Telling a business partner their approach isn't working.
  • Addressing a betrayal of trust.
  • A performance conversation that could end in termination.

Related

  • [[nonviolent-communication]] — the opening structure this composite runs.
  • [[active-listening]] — the receiving discipline this composite depends on.
  • [[conflict-mediation]] — the third-party version of this same challenge.

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

Install and usage instructions live in the source repository linked above.

Reviews

No reviews yet — be the first.

Versions

  • v0.1.0 Imported from the upstream source.