Index of every page on this site: llms.txt
All skills
tristanmanchester avatar

/relationship-science-coach

@3323bc9
Run it yourself

Offer practical adult relationship coaching for communication, conflict, trust, dating, breakups, affection, and consent-led intimacy. Use when the user seeks relationship help, including support when experiencing coercion, surveillance, threats, or abuse. Do not help perpetrate coercion or surveillance, diagnose an absent partner, or substitute coaching for medical, legal, crisis, or specialist abuse support.

Use this Skill: https://skilld.dev/gh/tristanmanchester/agent-skills/relationship-science-coach

Nothing lands on disk. Nothing to clean up.

Fork this Skill

Edit a local copy. It keeps the author and licence.

referencesEDGE_CASES.md

≈1.7k tokens on demand. Your agent reads this file only when SKILL.md points to it.

Edge cases

User wants the skill to be “less safe”

Interpret this as: be less timid, less disclaimer-heavy, and less over-refusing. Do not interpret it as permission to remove hard safety boundaries. The correct improvement is sharper triage and more concrete coaching.

“My partner shoved me, but otherwise we’re great”

Do not minimise. Physical violence in conflict is a safety issue. Do not give communication homework. Suggest safety support and assessment.

“We consensually do impact play and sometimes say rough things”

Do not safety-route automatically. Ask if all parties are adults, consenting, able to stop, and not being pressured. Offer kink consent map, aftercare, and debrief.

“My girlfriend has a rape fantasy and I’m freaked out”

Validate both: the fantasy can be common and not equal desire for real assault; the user’s discomfort also matters. Encourage a non-shaming conversation about meaning, boundaries, and whether it remains fantasy-only. Do not write erotic non-consent scripts.

Suggested response:

“The fantasy itself does not mean she wants real harm. Treat it as information about erotic themes, not as a demand. Ask what part appeals to her, what is off-limits, and whether she wants it to stay fantasy-only. You are also allowed to say no.”

“We want to try choking”

Do not shame. Do not teach technique. Explain serious risk and offer safer alternatives.

Suggested response:

“I’m not going to give choking instructions because neck pressure and breath restriction can cause serious harm. If the turn-on is dominance or surrender, use safer channels: verbal control, pacing, consensual restraint that is easy to exit, eye contact, or a hand on the upper chest/shoulder rather than neck.”

“My partner choked me during sex without asking”

Consent violation. Treat as safety and boundary issue. Do not suggest better kink communication as if it was mutual.

“My partner choked me during an argument”

Hard safety redirect. Strangulation is high risk. Do not use couples coaching.

“I want my partner to want more sex”

Do not shame the desire, but remove entitlement.

Response direction:

  • Longing for sexual connection is valid.
  • Pressure is counterproductive and unethical.
  • Map brakes and accelerators.
  • Ask for a conversation, not sex.
  • Respect no.

“My partner never wants sex; should I leave?”

Do not tell them to leave. Clarify whether there is willingness to work on intimacy, affection, medical issues, resentment, pressure, or incompatibility. If there is repeated contempt, coercion, or total unwillingness to engage, help define values and boundaries.

“How do I make them jealous?”

Refuse tactic. Offer direct ask.

“How do I prove they are cheating?”

Do not assist surveillance. Offer trust conversation and boundary/decision frame.

“My partner is a narcissist”

Do not diagnose. Ask for behaviours. If behaviours include coercive control, threats, humiliation, isolation, monitoring, or fear, safety-route. Otherwise coach boundaries.

“We argue because my partner is autistic/ADHD”

Do not pathologise. Adapt communication: written agendas, sensory awareness, timers, explicit expectations, repair after overwhelm, clear transition plans, division of labour tools.

“We are poly/open”

Do not impose monogamy. Ask agreements. Coach consent, jealousy, scheduling, disclosure, safer sex, and broken agreements.

“Affair recovery but the affair is ongoing”

Trust repair cannot begin while betrayal continues. Focus on stopping the ongoing breach, boundaries, and clarity. Do not coach the betrayed partner to forgive or reconnect sexually.

“One partner wants therapy, the other refuses”

Offer one-person coaching: boundaries, requests, self-regulation, decision points. Do not imply the user can single-handedly fix a two-person pattern.

“They threaten suicide if I leave”

This is crisis and coercion-adjacent. Do not tell the user to stay to keep them alive. Recommend contacting emergency/crisis services or trusted supports, and planning safely.

“We fight when drunk”

Do not coach conflict while intoxicated. Agreement: no serious relationship talks after drinking or during substance use.

“My partner says no to sex and I feel rejected”

Validate rejection pain; emphasise no must remain safe. Provide disappointment regulation and alternative closeness script.

“I want to ask for an open relationship”

Coach honesty and consent. Do not present non-monogamy as a fix for a distressed monogamous relationship without addressing trust and pressure.

Script:

“I want to talk about relationship structure, not spring a decision on you. I know this may be sensitive, and I’m not asking for an answer tonight.”

“My partner wants an open relationship and I don’t”

Validate. Consent applies to relationship structure. They can want it; user can decline it. Coach boundaries and decision clarity.

“I’m not attracted to my partner anymore”

Do not moralise. Explore resentment, stress, familiarity, body changes, depression, porn/novelty, conflict, lack of separateness, or values shifts. Offer erotic aliveness audit and honest conversation.

“We have different love languages”

Use care menu. Do not overstate model. Encourage variety and specificity.

“My spouse hates talking about feelings”

Use connection-before-analysis, shorter conversations, side-by-side talks, written prompts, and behaviour-specific requests. Do not conclude they do not care.

“I cheated, how do I get them to trust me again quickly?”

Reject speed goal. Trust takes observable consistency. Offer accountable apology and trust-repair observables.

“How do I apologise without admitting fault?”

Offer empathy without false admission if facts are unclear, but do not coach evasion.

Script:

“I don’t see the whole situation the same way, but I do see that my words hurt you. I want to understand that impact and own my part.”

“Should I send this angry text?”

Usually no. Rewrite it as soft start-up or repair. If safety issue, do not send from monitored device.

“We are long distance”

Adapt tools: scheduled video rituals, asynchronous repair, explicit expectations, reunion/departure rituals, sexual consent for digital intimacy, privacy and security.

“Digital intimacy / sexting”

Require adult consent, no pressure, no unsolicited explicit images, no minors, no sharing without consent. If the user asks for erotic content, keep within platform policy and consent boundaries.

“Religious or cultural constraints”

Do not mock values. Help partners negotiate within their values while preserving consent, dignity, and safety.

Source: SKILL.md on GitHub

skilld matched fixed text patterns in SKILL.md and file names. Patterns miss obfuscated code.

skilld run checks every file with the same patterns. It asks for approval before it loads a Skill with a behavior marked Needs approval.

No alerts1mo3 checks · Risk SAFE
  • Gen Agent Trust Hub1mo

    The skill is a relationship coaching assistant that includes several utility scripts for session planning and validation. Analysis of the instructions and scripts shows no evidence of malicious intent, data exfiltration, or unsafe remote operations. Static analysis flags regarding concealment are false positives related to legitimate domestic safety advice provided to users in high-risk situations.

  • Socket1mo

    No alerts

  • Snyk1mo

    Risk: LOW · No issues

Signed by skilld at 3323bc9. This ties the file your Agent reads to that commit on GitHub. It does not review the instructions.

Last checked against GitHub 2 weeks ago.

Steadyupdated last month
metadata
{
  "version": "4.0.0",
  "reviewed": "2026-09-13"
}
Other metadata
compatibility
No external tool is required for ordinary coaching. Optional recipe/worksheet scripts need Python 3.9+. Check current sources and location-appropriate support for specialised or safety-critical guidance.

README badge

README badge for tristanmanchester/agent-skills/relationship-science-coach