Skip to content

paultyng/skill-issue / skills/match-voice

match-voice

Verified

Use when asked to rewrite or format text in the user's voice, "make this sound like me", "match my voice", "de-stiffen this", when text "reads too corporate / businessy / stiff", to "strip the em-dashes / agent typography", or to draft a Slack announcement, PR body, or doc in the user's voice. Also /match-voice.

FileKindSize
SKILL.mdskill md5.3 KB
reference.mdmarkdown3.4 KB

reference.md

View on GitHub
# Voice corpus

Verbatim anchors for the `match-voice` profile, from a 6-month analysis (Slack, Claude Code sessions, GitHub PR/review prose, Notion docs). Use them to calibrate tone and word choice; do not copy them literally into output.

Some quotes below were originally typed lowercase with dropped apostrophes (Slack/chat). **That casing is incidental — not the voice.** The signal is the tone, bluntness, and phrasing. Output is always properly capitalized; the personality survives the jump.

## Contents
- Voice anchors
- Signature moves
- Anti-voice (what generated text does that the user does not)

## Voice anchors

Dry, blunt, self-deprecating, plain. Casing here is the original surface; keep the edge, normalize the case.

- "i just test local now and yolo merge things, what could go wrong!" (the dry fatalism, properly cased: "We test locally and merge on green.")
- "we have IP rate limiting in the WAF, is there a reason to re-implement it in the service here?" (questions a decision instead of asserting)
- "here is my WIP draft, pretty early for comments, but i will begrudgingly accept them if you must"
- "easy for me to wildly speculate about things that don't exist though!" (self-deprecating aside)
- "Auto-management was paternalistic: idle sessions were getting killed mid-context; RSS-limit was a defensive cap against an upstream memory leak that should be upstream's problem."
- "Silence is not approval. A silent stakeholder is a kick-back reason."
- "Human gates are fixed; everything else is a dial."
- "If an effort isn't worth specifying, it's not worth doing." (the "this is a feature, not a bug" move)
- "These are deferred, not abandoned."
- "Predictability comes from ticket count, not point velocity."
- "Closing — going with manual enforcement, clearer guidance in RELEASING.md. No programmatic check needed; reviewers + the runbook handle it."

## Signature moves

- Conclusion first, justification after (or omitted).
- Uncertainty labeled explicitly: "unsure", "not sure", "open question" — a flag, not a hedge. Separates "I know" from "I suspect".
- Parenthetical asides; bullets over prose; bold for the load-bearing phrase.
- Plain short verbs: wire, gate, cap, shim, sweep, surface, flesh out, lock, pull.
- Dry, wry, lightly self-deprecating. Collaborative "we".
- Recurring tags in code/docs: "follow-up", "out of scope" / "not in this PR", "read-only", "portable", "pre-existing".

## Anti-voice

Observed in AI-generated sections on the user's own pages — the register to avoid:

- "emphasized the importance of", "highlighted", "stressed", "The team discussed..."
- "data-driven decision-making framework", "leverage", "synergize", "actionable insights", "moving the needle", "robust", "seamless"
- Enthusiasm openers ("Excited to announce...", "This PR adds exciting...").
- Hedge pileups ("I think this might possibly be a concern").
- Performed sincerity: "honest" / "honestly" / "to be honest" / "I'll be honest" / "frankly" / "let's be real". The user labels doubt ("unsure", "open question") instead of announcing candor.
- Common AI filler: "It's worth noting", "That said" / "That being said", "Here's the thing", "not just X, but Y", "deep dive", "delve", "unpack", "landscape", "realm", "crucial", "vital".
- Closing pleasantries ("Let me know if you have questions!").
- Generic headers (Overview, Background, Introduction).
- Em-dashes, smart quotes, unicode arrows/bullets — the user types none of these.
reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis.Stars: 471kUpdated Claude CodeCodexVerifiedInstall
diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness.Stars: 322kUpdated Claude CodeCodexVerifiedInstall
dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependenciesStars: 322kUpdated Claude CodeCodexVerifiedInstall
executing-plansobra/superpowersUse when executing an implementation plan in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is availableStars: 322kUpdated Claude CodeCodexVerifiedInstall