skills/professional-communication/SKILL.md
Use when drafting high-stakes business messages (layoffs, escalations, disagreements with leadership, budget requests, bad news upward). Use when the user's communication problem is political, cross-cultural, or audience-calibration -- not just 'make this sound professional.' NEVER for casual messages, basic email formatting, or grammar fixes.
npx skillsauth add sharkitect-solutions/sharkitect-claude-toolkit professional-communicationInstall this skill globally with one command. Works with Claude Code, Cursor, and Windsurf.
3 of 9 scanners reported clean
Some scanners were skipped, did not run, or reported a non-clean status. Review each row below.
Before writing anything, diagnose the ACTUAL problem. Most people ask for "help writing an email" when the real issue is upstream.
What is the user struggling with?
|
|-- "I don't know what to say"
| --> MESSAGE STRUCTURE problem. Use the Pyramid or BLUF framework below.
|
|-- "I don't know how to say it without causing problems"
| --> POLITICAL problem. Go to Political Communication section.
|
|-- "They keep misunderstanding me"
| --> AUDIENCE CALIBRATION problem. Check the Seniority Ladder.
| --> Or MEDIUM MISMATCH. Check the Medium Decision Tree.
|
|-- "I need to deliver bad news"
| --> HIGH-STAKES MESSAGE. Go to The 5 Hardest Messages.
|
|-- "This person is in another country/culture and things keep going wrong"
| --> CROSS-CULTURAL problem. Go to Cross-Cultural Traps.
|
|-- "I need to get something approved"
| --> PRE-WIRE problem. Go to The Pre-Wire Technique.
|
|-- "I just need it to sound more professional"
| --> This skill is probably overkill. Claude already knows tone adjustment.
| --> Apply BLUF + cut filler. Done.
| What the user says | What they actually need | |---|---| | "Help me write a professional email" | Diagnose whether the problem is structure, politics, audience, or medium first | | "Make this sound nicer" | Identify what is politically dangerous in their draft and neutralize it | | "I need to escalate this" | A message that escalates the ISSUE without escalating the CONFLICT | | "How do I tell my boss they're wrong?" | A face-saving reframe that lets the boss arrive at the right answer | | "I need to send bad news" | A message structured so the reader processes facts before emotions | | "Can you write this for multiple audiences?" | Separate messages calibrated per audience, not one compromise message | | "Help me follow up -- they're ignoring me" | Diagnose why they are ignoring (unclear ask, wrong person, no incentive) | | "I need to be more concise" | Their real problem is burying the lead -- fix structure, not word count |
The biggest communication mistake is talking to an executive like a peer or to a peer like an executive.
IC to IC: Lead with the problem, share technical context, ask for help directly. Informality is fine. "Hey, the auth middleware is dropping sessions after 30 min -- have you seen this? Here's my debug log."
IC to Manager: Lead with impact and what you need from them. Skip implementation details unless asked. "The auth bug is blocking the release. I need 2 days to fix it, which pushes the deadline to Friday. Want me to proceed or should we ship without the fix?"
IC/Manager to Director+: Lead with business impact and a recommendation. They want to make a decision, not understand your process. Use the Pyramid Principle: conclusion first, supporting evidence in layers, details only if asked.
Anyone to Executive (VP/C-level): Maximum 3 sentences for the core message. The "so what" chain must end at revenue, risk, or customer impact. Everything else is noise. Pre-wire the decision before the meeting if possible.
Downward (to reports): Lead with context and reasoning, not just the decision. People execute better when they understand why. But do not over-explain -- it signals you do not trust them to handle it.
Most communication failures are medium mismatches, not message failures.
Is the message emotionally charged or politically sensitive?
|-- Yes --> Voice/video call, then summarize in writing
|-- No:
|
Does the message require back-and-forth to reach a decision?
|-- Yes, 2-3 exchanges --> Chat/Slack thread
|-- Yes, complex --> Schedule a 15-min call
|-- No:
|
Does it need to be findable in 30 days?
|-- Yes --> Email or shared document
|-- No --> Chat is fine
Key insight: The medium IS part of the message. Sending a Slack DM about a layoff communicates "this doesn't matter." Sending a formal email about a lunch plan communicates "I'm difficult to work with."
The goal is never to prove you are right. The goal is to help them arrive at the better answer while preserving their authority.
The Reframe Technique: Do not say "I disagree" or "I think you're wrong." Instead, add a constraint they may not have considered:
BAD: "I don't think we should use vendor X. Their uptime is terrible." GOOD: "I looked into vendor X -- they hit most of our requirements. One thing I want to flag: their SLA is 99.5%, and our compliance team requires 99.9%. Want me to check if they offer a higher tier, or should I pull alternatives?"
This works because: (1) you show you did the work, (2) you introduce a fact, not an opinion, (3) you offer them the choice rather than making it for them.
Escalation is not tattling. It is transferring a decision to someone with the authority to make it. But it feels like tattling if you do it wrong.
Before escalating, always:
Escalation message structure:
Bad news must travel fast, with facts, and with a plan.
Structure: SIP (Situation - Impact - Plan)
BAD: "Hey, wanted to give you a heads up that there might be some delays with the project. We ran into some issues and are working through them. Will keep you posted."
GOOD: "The payment integration failed load testing -- it drops transactions above 500/min. This blocks our March 15 launch. We have two options: (A) add a queue layer, which takes 5 days and pushes launch to March 20, or (B) launch with a 500/min cap and patch in sprint 2. I recommend option A. Can we discuss today?"
The most important communication in business happens BEFORE the meeting, not during it.
Pre-wiring means individually socializing your proposal with key stakeholders before presenting it to the group. By the time the meeting starts, you already know who supports you, who has concerns, and what the objections will be.
How to pre-wire:
When NOT to pre-wire: When transparency requires real-time group deliberation (ethics decisions, layoff discussions, incident reviews).
These cause real damage on distributed teams and are rarely taught:
Direct vs. Indirect cultures: In the US, Netherlands, Israel, Germany -- directness is expected. "This won't work because..." is professional. In Japan, Korea, many Southeast Asian cultures, the UK -- indirectness is the norm. "This is interesting, but perhaps we could consider..." means "no." If you miss this, you will think agreement happened when it did not.
Silence in meetings: In some cultures, silence means processing or respect. In US meeting culture, silence means agreement or disengagement. If your distributed team has members from high-context cultures, explicitly ask: "I want to make sure everyone has had a chance to share concerns. [Name], what are your thoughts?"
"Yes" does not always mean yes: In many cultures, saying "no" directly to a superior or client is unacceptable. "Yes, we will try" may mean "this is impossible but I cannot say that." Follow up with: "What would you need to make this work by Friday?" If the answer is vague, the real answer is probably no.
Written tone across cultures: Exclamation points, emojis, and casual greetings that read as friendly in American English can read as unprofessional in German or Japanese business contexts. When in doubt, err formal for the first 3-4 exchanges, then mirror the other person's tone.
Lead with the decision (do not bury it). State the reason honestly. Explain what happens next for the affected person. Do not over-explain or apologize excessively -- it centers your discomfort instead of their needs.
Key rules: Be specific about severance, benefits continuation, and timeline. Never say "this was a difficult decision" without concrete next steps. Never blame the person ("performance-based" when it is actually a reorg). Have the conversation live first, then send the written follow-up.
Acknowledge the team's work before explaining the decision. State the business reason clearly. Address what happens to the people (reassignment, not limbo). Do not frame cancellation as "pivoting" or "refocusing" -- the team knows what happened, and euphemisms destroy trust.
Use the Reframe Technique above. Never frame it as "you're wrong." Always frame it as "here's a constraint/data point that might change the calculus." Offer two paths, including theirs, with honest tradeoffs. Make it easy for them to change course without losing face.
Executives approve budgets based on ROI and risk, not on how hard you are working. Structure: (1) What you will deliver with the resources, (2) What happens without them (be specific -- "we miss the Q3 target" not "things will be harder"), (3) The specific ask in dollars and headcount, (4) Timeline to ROI.
Never lead with "we need more people." Lead with what the business gets.
Separate observation from interpretation. "You missed 3 of the last 5 deadlines" is observation. "You don't care about the team" is interpretation. State the pattern, state the impact, ask for their perspective before prescribing solutions. Document everything. Never deliver this over email first -- it requires a conversation, with the written version as follow-up.
BEFORE (actual message sent by a senior engineer):
Subject: Concerns about the new architecture
Hi team,
I've been thinking about the new microservices proposal and I have some concerns. I think we might be over-engineering this. The current monolith works fine for our scale and adding microservices will create a lot of operational overhead that we don't have the team to support. I've seen this pattern before at my last company and it didn't go well. I think we should reconsider.
Also, I noticed that the proposal doesn't address data consistency across services, which is going to be a major issue.
Let me know what you think.
What went wrong:
AFTER (rewritten for effectiveness):
Subject: Architecture proposal -- data consistency question
Hi [Architect name],
I reviewed the microservices proposal. The service boundaries look clean. One area I think needs more detail before we commit: data consistency across services. Specifically:
- Orders and inventory would span two services. How do we handle a scenario where an order is placed but inventory update fails?
- The proposal doesn't mention saga patterns or eventual consistency tradeoffs. Should we add a section on this?
I also want to flag an operational concern: we currently have 3 engineers on infra. Microservices typically need at least 5 for on-call coverage across services. Should we factor hiring into the timeline?
Happy to pair on the data consistency section if that would help.
What changed: Direct to the right person. Leads with the technical gap. Asks questions instead of asserting opinions. Offers to help rather than just criticize. Quantifies the operational concern instead of vague "overhead."
development
When the user wants help with paid advertising campaigns on Google Ads, Meta (Facebook/Instagram), LinkedIn, Twitter/X, or other ad platforms. Also use when the user mentions 'PPC,' 'paid media,' 'ad copy,' 'ad creative,' 'ROAS,' 'CPA,' 'ad campaign,' 'retargeting,' or 'audience targeting.' This skill covers campaign strategy, ad creation, audience targeting, and optimization.
testing
--- name: using-sharkitect-methodology description: Use when starting any conversation in a Sharkitect workspace OR before any task involving NEW pricing, positioning, proposal, strategy, plan-execution, or schema-design work — mandates invocation of Sharkitect-specific methodology skills (pricing-strategy, marketing-strategy-pmm, smb-cfo, hq-revenue-ops, executing-plans, brainstorming) under the same anti-rationalization discipline as using-superpowers. Documentation has failed 4 times across H
testing
Use when user says 'end session', 'wrap up', 'stop for the day', 'done for today', 'close out', 'save session', 'wrapping up', or invokes /end-session. Runs the full 9-step end-of-session protocol: resource audit, MEMORY.md update, lessons capture, plan status, pending items, workspace checklist, .tmp/ audit, git commit+push, Supabase brain sync, session brief, summary. Final step schedules a detached self-kill of the current session ONLY (3s delay) so the window closes cleanly. Other claude.exe processes (active workspaces) are NOT touched -- orphan cleanup is handled separately by Claude-Orphan-Cleanup-Hourly with proper age safeguards. Do NOT use for: mid-session quick saves (use session-checkpoint), skill syncing (use sync-skills.py), brain memory queries (use supabase-sync.py pull), document freshness reviews (use document-lifecycle), resource gap detection (use resource-auditor).
testing
Remove signs of AI-generated writing from text. Use when editing or reviewing text to make it sound more natural and human-written. Based on Wikipedia's comprehensive "Signs of AI writing" guide. Detects and fixes patterns including: inflated symbolism, promotional language, superficial -ing analyses, vague attributions, em dash overuse, rule of three, AI vocabulary words, passive voice, negative parallelisms, and filler phrases.