phuryn

The durable documentation set that makes an AI-built (vibe-coded) app reviewable before shipping. A small core every app needs — architecture, user/permission flows, permissions, variables/secrets, and a test-coverage map — plus conditional docs added only…

The method for finding the gap between what a system is supposed to do and what the code actually does — the class of bug generic scanners miss because they have no model of intent. Defines what counts as documented intent, what counts as implementation…

Create a lean, decision-ready Product Requirements Document with an 8-section template: summary, contacts, background, objective, market segments, value propositions (6-part JTBD template), solution, and release. Lean by default (about one to two pages), with…

Review code for actionable defects. Correctness is the core; performance and security are optional sub-cases of the same engine. Anchors on agreements between participants across a boundary, forces a violating execution, and refutes every candidate before…

Red-team a PRD, roadmap, or strategy by attacking its load-bearing assumptions before reality does. Steelmans then attacks each claim, ranks failure modes by impact × likelihood × cheapness-to-test, and returns the cheapest test and kill criteria for each.…

Analyze A/B test results with statistical significance, sample size validation, confidence intervals, and ship/extend/stop recommendations. Use when evaluating experiment results, checking if a test reached significance, interpreting split test data, or…

Perform cohort analysis on user engagement data — retention curves, feature adoption trends, and segment-level insights. Use when analyzing user retention by cohort, studying feature adoption over time, investigating churn patterns, or identifying engagement…

Generate SQL queries from natural language descriptions. Supports BigQuery, PostgreSQL, MySQL, and other dialects. Reads database schemas from uploaded diagrams or documentation. Use when writing SQL, building data reports, exploring databases, or translating…

The house writing style. Consult before finalizing any reply.

The team's house writing style. Consult before finalizing any reply.

The team's house writing style. Consult before finalizing any reply.

Use whenever you write a report or summary. It defines the required report structure. Load it before writing and follow it exactly.

Use this on EVERY answer the team gives. It defines the team's mandatory house style for closing a reply. Load it before finalizing any response and apply it.

Deploy, run, list, or tear down the agents in this repo's .managed-agents/ folder on Anthropic Managed Agents. Use when the user says "deploy my agents", "ship this agent to the cloud", "run the deployed agent", or asks about managed-agent deployment.

Use whenever you state a fact. End the answer with a short Sources list mapping each claim to a file, URL, or reasoning. Shared across agents.

Apply the Acme house style to any prose - short sentences, no hype, lead with the answer. Shared across agents.

Edit AI-sounding text so it reads like a careful person wrote it, without changing what it says or commits to. Built for work writing where the wording matters: sales and marketing copy, support replies, HR and legal letters, emails, docs, release notes,…

Edit AI-sounding text so it reads like a careful person wrote it, without changing what it says. Two modes. Rewrite: edit the text you are given (posts, emails, docs, release notes, replies, landing pages). Interview: when a draft is AI-composed and thin, ask…

Edit AI-sounding text so it reads like a careful person wrote it, without changing what it says. Two modes. Rewrite: edit the text you are given (posts, emails, docs, release notes, replies, landing pages). Interview: when a draft is AI-composed and thin, ask…

Edit AI-sounding text so it reads like a careful person wrote it, without changing what it says. Two modes. Rewrite: edit the text you are given (posts, emails, docs, release notes, replies, landing pages). Interview: when a draft is AI-composed and thin, ask…

Edit AI-sounding text so it reads like a careful person wrote it, without changing what it says. Two modes. Rewrite: edit the text you are given (posts, emails, docs, release notes, replies, landing pages). Interview: when a draft is AI-composed and thin, ask…

Edit AI-sounding text so it reads like a careful person wrote it, without changing what it says. Two modes. Rewrite: edit the text you are given (posts, emails, docs, release notes, replies, landing pages). Interview: when a draft is AI-composed and thin, ask…

Edit AI-sounding text so it reads like a careful person wrote it, without changing what it says. Two modes. Rewrite: edit the text you are given (posts, emails, docs, release notes, replies, landing pages). Interview: when a draft is AI-composed and thin, ask…

Edit AI-sounding text so it reads like a careful person wrote it, without changing what it says. Two modes. Rewrite: edit the text you are given (posts, emails, docs, release notes, replies, landing pages). Interview: when a draft is AI-composed and thin, ask…

Use OpenAI Codex CLI as a peer collaboration partner — same standing as Claude, not a one-shot tool. Bring Codex in EARLY (ideation, framing decisions, stuck moments), not only for late peer-review. Hand it strategic context (target, research, draft,…

DON'T USE THAT SKILL. IT'S JUST AN EXAMPLE. Summarize today's Gmail inbox by category with anonymized senders and subjects, then surface the top 3 emails to focus on next. Run only when the user types /email-summary.

Scan Gmail (default: today's inbox), produce an anonymized summary with action points, then export each email to Linear as an issue with a sensible priority. No personal details — names, email addresses, or company identifiers — ever leave the summary or…

Spawn a new Claude Code session with remote control enabled (default: a detached background session — via screen on macOS/Linux, via a minimized PowerShell Start-Process on Windows — fully remote-safe, no approvals/windows), so it appears in the mobile Claude…

Initialize a PM Brain — a markdown-native second brain for a product operator (PM, product lead, founder, or anyone accountable for one product or initiative) doing judgment-heavy work with scattered inputs. Detects greenfield vs. migration mode, runs a…

Review a Python solution against a markdown spec, run it, and report bugs.

Edit AI-sounding text so it reads like a careful person wrote it, without changing what it says or commits to. Built for work writing where the wording matters: sales and marketing copy, support replies, HR and legal letters, emails, docs, release notes,…

Embed Claude agents in a web app with the Claude Agent SDK. One server-side wrapper runs any agent kept as a directory under .claude/agents/<name>/ (its AGENTS.md, knowledge files, skills and MCP servers), isolated from the repo and the machine, streamed to a…