AI Agent Workflow Collection

Use when a user pastes, attaches, or points at code and asks for opinionated critique — especially brutal, savage, funny, harsh, honest, or merciless feedback. Strong triggers: "review this," "what's wrong with," "tell me everything wrong," "brutal/savage feedback," "roast," "judge," "audit," "grade," "rip apart," "don't hold back," "is this bad," "embarrassing," profanity about their own code ("garbage/dogshit/trash"), diff reviews before committing, "should I push this," comparing two snippets, whole-project grading, or meta-reviews of SKILL.md files. User wants findings with severity ratings, concrete fixes, and personality — not a dry lint report. Trigger even when "roast" or "blunt cake" never appear; the signal is subjective code critique with attitude. Do NOT trigger for: writing new code, adding features, translating languages, running formatters/linters, writing tests, explaining how code works, or library recommendations.

Use when a user shares web code (HTML, JSX, TSX, Vue, Svelte) and wants to know if it's accessible — keyboard usable, screen reader compatible, WCAG compliant, or ready to ship from an a11y standpoint. The trigger is code + an accessibility question: "does this pass WCAG", "check this for a11y", "is this keyboard accessible", "audit before I ship", "compliance report for a client". Does NOT apply when the user wants to write new accessible components, understand WCAG concepts abstractly, find accessibility libraries, or write accessibility tests.

Start a focus timer for deep work sessions. Tracks elapsed time and optionally reminds about breaks. Use when the user wants to focus, start a timer, or do a pomodoro session.

Check real rate limit usage from the Anthropic API — session (5-hour), weekly, and Opus utilization. Use when asking about rate limits, pacing, token usage, "how much runway do I have", "should I keep working", "am I burning too fast", "will I hit my limit", or wanting a forecast of remaining subscription budget.

Check if you're in Anthropic's peak throttling window. Shows status for any timezone. Use when asking about peak hours, rate limits, or best time to work.

Check your current session's health — real context usage from JSONL token data, compaction proximity, session weight, and rate limit status. Use when asking about session size, compaction, context window, rate limits, "how big is my session", "how close to compaction", or "should I checkpoint".

Check how long the current session has been running. Use when asking about session length, duration, "how long have I been at this", "how long have I been working", whether it's time to checkpoint, or any time-since-start question.

Time-aware assistant for Claude Code session management. Use when the user asks about time, session health, peak hours, rate limits, pacing, work patterns, focus mode, or wants a daily briefing. Also use when the user says "father time" or asks about their session duration, context usage, or compaction status.