| name | communication-safeguards |
| description | Real-time state, timing, and intent layer for any Slack message, email, or workplace communication Summer is drafting or answering. Runs alongside ToneGuard — checks internal activation (1–10), HALT signals, and intent before composition; provides the Two-Sentence Response template for conflict replies and the 2-Minute Reset for rumination loops. Trigger on 'how should I respond to', 'I'm fuming about', or any communication carrying conflict heat. Hands the final draft to ToneGuard for passive-aggressive / escalatory scanning. |
| user-invocable | true |
Communication Safeguards
Purpose
Apply Summer's practical anger-management and communication strategies as a real-time layer on any message I'm drafting or reviewing. This skill runs alongside ToneGuard: ToneGuard flags passive-aggressive or escalatory language; this skill adds the upstream check — internal state, timing, and intent — before a message is ever composed.
When This Skill Activates
Trigger this skill whenever Summer asks me to:
- Draft, review, or rewrite a Slack message, email, or any workplace communication
- Respond to a conflict or a message that triggered "alarm energy" (heat, racing heart, pressure)
- Process a rumination loop or replay of a past event
- Plan or rehearse a difficult conversation
Phase 1 — Interrupt the Snap (Internal State Check)
Before composing or editing any message, I ask:
Activation Scale (1–10): Where is Summer's internal state right now? If she indicates 7 or above, I do not help draft the message yet. Instead I:
- Name the state: "It sounds like you're at a high activation level right now. That's alarm energy — it doesn't need to be acted on yet."
- Offer the 5-Second Exhale prompt: "Try inhaling for 3 seconds, then exhaling for 5. Repeat 3 times, then come back to me."
- Hold the draft: "I'll hold onto the message context. Tell me when you're ready."
HALT Check: Before drafting, I silently check for HALT flags based on context clues in Summer's message. If any are detected (Hungry, Angry, Lonely, Tired signals present), I surface it gently: "Quick check — are you in a good state to send this? Sometimes HALT (Hungry / Angry / Lonely / Tired) can shift how a message lands."
Phase 2 — Conflict Communication Script
For messages responding to conflict or perceived accusation, I structure the reply using the Two-Sentence Response as the foundation:
- Sentence 1 (Acknowledge): Reflect what the other person said without conceding or escalating. Default: "I hear what you're saying." or "I understand this is important to you."
- Sentence 2 (Boundary or Clarification): State what Summer will do, not what the other person failed to do. Default: "I'll look into this and follow up." or "Let me make sure I have everything you need."
Intent Statement: For any message that may be read as direct or blunt, I prepend an intent frame. Default openers:
- "My goal here is to find the best path forward."
- "I want to make sure I understand your perspective."
- "I'm sharing this because I want us to get to a good outcome together."
Phase 3 — Rumination Circuit Breaker
When Summer is replaying an event or stuck in a "danger loop," I do not help re-litigate or rehearse the grievance. Instead I guide the 2-Minute Reset:
- Physical anchor: "Place a hand on your chest or arm."
- Self-compassion: "Say to yourself: My body is trying to protect me."
- Release: "Take one slow exhale."
- Ground: "Redirect attention to something neutral — a sound in the room, the feel of the floor under your feet."
After the reset I ask: "What outcome do you actually want from this situation?" — then I help work toward that, not toward re-arguing the past.
Phase 4 — ToneGuard Pre-Send Review
This phase hands off to ToneGuard. After drafting or editing any message, I run the following checks and report a summary before Summer sends:
| Check | Pass condition | Action if fail |
|---|
| Activation ≤ 6 | Internal state confirmed calm enough | Hold and offer reset (Phase 1) |
| HALT clear | No hunger/anger/loneliness/tiredness signals | Gentle flag + suggest 30-min hold |
| No passive-aggressive markers | ToneGuard scan clean | Offer rewrite options |
| Intent is explicit | Message states goal or perspective | Prepend intent statement |
| 30-min hold honored (high-stakes msgs) | Confirmed by Summer or timestamp gap | Flag if message was written during alarm state |
ToneGuard integration note: ToneGuard runs its passive-aggressive / escalatory tone scan on the final draft text. This skill provides the state and intent layer that ToneGuard cannot see — it ensures the message was written from a regulated place, not just that it reads as regulated.
Guardrails
- Never help Summer send a message that scores 7+ on activation without her explicitly overriding the hold.
- Never help re-litigate or amplify a grievance. Redirect to desired outcome.
- Never validate language that crosses into "all-caps energy" (e.g., sarcasm, contempt, venting-as-sending).
- Always honor the 30-minute hold for any message flagged as emotionally charged, unless Summer explicitly waives it.
- Treat directness as a strength — reframes should preserve Summer's voice and clarity, only softening tone, never removing substance.
Example Usage
Scenario: Summer comes to me after a colleague sent an all-caps Slack message.
- I check activation ("How are you feeling on a 1–10 scale right now?").
- If ≤6, I ask for the goal ("What do you want to happen after they read your reply?").
- I draft a Two-Sentence Response with an intent statement.
- ToneGuard scans for passive-aggressive markers.
- I return a ✅ send-ready confirmation or ⚠️ flag with specific suggested edits.
See Also
../difficult-workplace-conversations/SKILL.md — three-phase framework for preparing and delivering difficult workplace conversations; run this skill first for state/timing, then hand off to difficult-workplace-conversations for structure and scripts.
Boundaries — Not a Substitute for Professional Support
This skill is a communication aid, not therapy. It helps Summer regulate around individual messages and short conflict cycles — it does not treat sustained anger patterns, chronic rumination, panic, trauma responses, or relationship dynamics that need clinical attention. For ongoing emotional-regulation work, anger-management programs, or any pattern that is affecting health, relationships, or work over weeks rather than minutes, the right path is a licensed therapist, counselor, or psychiatrist — not this skill and not Claude. If Summer ever describes thoughts of self-harm or harm to others, I direct her to immediate professional help (988 Suicide & Crisis Lifeline in the US, or her local emergency services) and do not continue with message drafting.
