Irfad Imtiaz

When working on sensitive projects, proprietary code, or systems where context leakage is a concern. Use when the user says "protect this," "sensitive project," "don't leak," "sanitize output," "clean context," "proprietary," "confidential," or when working across multiple projects where context bleed between them could be problematic. Implements defensive patterns to prevent sensitive context from leaking into outputs, commits, or cross-project contamination.

When ending a session, consolidating memory, or when memory files feel stale or contradictory. Use when the user says "dream," "consolidate memory," "clean up memory," "merge notes," "memory maintenance," "what do you remember," "update your knowledge," "session end," or when memory/context files have accumulated noise over multiple sessions. Performs cross-session memory synthesis inspired by KAIROS autoDream.

When the user needs to perform the same operation across multiple files, components, or entities. Use when the user says "do this to all," "batch," "bulk update," "across all files," "every component," "mass refactor," "rename everywhere," "update all," "migrate all," or when a task involves applying a consistent pattern to 3+ targets. Orchestrates parallel processing of repetitive tasks with consistency guarantees.

When the user wants to find bugs, test edge cases, stress-test code, or perform adversarial analysis of their codebase. Use when the user says "find bugs," "break this," "what could go wrong," "red team," "hunt bugs," "adversarial test," "edge cases," "stress test," "what did I miss," "pen test this logic," or before any major release. Operates with red team psychology — assumes everything is broken until proven otherwise.

When the user wants to deeply understand a codebase, map its architecture, or build a persistent mental model for future sessions. Use when the user says "explain this codebase," "map the architecture," "codebase insight," "understand this repo," "what does this project do," "how does this work," "analyze the codebase," or when starting work on an unfamiliar project. Produces a persistent insight file that survives across sessions.

When the user wants a thorough code review of changes, a PR, a file, or a feature branch. Use when the user says "review this," "code review," "check my code," "what's wrong with this," "review my PR," "look over this," "critique this," "review before merge," or when submitting code for quality assessment. Performs multi-pass review covering correctness, security, performance, maintainability, and edge cases.

When conversation context is getting long, when approaching token limits, when the user says "compress," "summarize context," "too long," "running out of context," "compact," or when you notice degraded performance due to context length. Implements a 3-layer compression pipeline (MicroCompact, AutoCompact, Full Compact).

When the user has a complex task that would benefit from multiple agents working in parallel or in sequence. Use when the user says "coordinate," "multi-agent," "parallelize this," "split this up," "delegate," "swarm," "fan out," or when a task naturally decomposes into 3+ independent subtasks. Orchestrates complex workflows by breaking work into units, dispatching agents, merging results, and resolving conflicts.

When the user needs accurate effort estimates for tasks, features, or projects. Use when the user says "how long will this take," "effort estimate," "estimate this," "scope this," "size this," "time estimate," "sprint planning," or when planning work. Unlike generic estimates, this skill calibrates against YOUR actual velocity — past estimates vs actuals — so estimates get more accurate over time.

When the user shows signs of frustration, impatience, or dissatisfaction. Also use proactively to calibrate tone. Triggers on phrases like "this isn't working," "why does this keep," "I already told you," "again?!", "just do it," "stop," "wrong," "that's not what I asked," "are you even listening," or any swearing/ALL CAPS. Detects frustration signals and fundamentally changes approach — not just tone, but strategy.

When starting or ending a work session, or when context files need maintenance. Use when the user says "boot up," "morning briefing," "end session," "session review," "what changed," "sync state," "kairos," "dream," "consolidate," "session start," or at the natural beginning/end of any significant work session. Implements structured session lifecycle with intelligent context loading, state diffing, and memory consolidation.

When the user wants to generate documentation from their codebase. Use when the user says "document this," "write docs," "generate README," "API docs," "architecture docs," "explain this codebase," "onboard someone," "write a guide," or when a project lacks documentation. Auto-generates comprehensive documentation by reading and analyzing the actual codebase.

When the user has PR review feedback to address, CI failures to fix, or wants to automatically resolve common PR issues. Use when the user says "fix the PR," "address review comments," "CI is failing," "fix the build," "autofix," "pr feedback," "resolve comments," "fix merge conflicts," or when a PR needs updates before merge. Reads the review comments, diagnoses issues, and applies fixes.

Use proactively after completing any task to suggest logical next steps. Also use when the user seems unsure what to do next, says "what now," "what else," "what should I do," "what's next," "continue," or pauses after a milestone. Anticipates what the user likely needs next based on context and proactively offers actionable suggestions.

When the user wants a security audit, vulnerability assessment, or security hardening of their codebase. Use when the user says "security review," "security audit," "check for vulnerabilities," "is this secure," "pen test," "OWASP," "find security issues," "harden this," "check for leaks," "secrets scan," or before deploying any user-facing application. Performs a systematic security audit covering OWASP Top 10, secrets scanning, dependency audit, auth review, and infrastructure security.

When ending a work session and wanting to capture what was accomplished, what was learned, and what to do next. Use when the user says "session review," "what did we do," "wrap up," "summarize session," "weekly review," "retrospective," "what happened today," "debrief," or at natural session endings. Reviews the session transcript, extracts learnings, identifies patterns, and produces an actionable summary.

When you or the user are going in circles, stuck on a problem, or repeating failed approaches. Use when the user says "I'm stuck," "this isn't working," "we keep going in circles," "try something different," "nothing works," "help," or when you detect that you've attempted the same category of solution 2+ times without success. Forces a structured step-back, reframe, and fresh approach.

When the user wants to review past decisions, understand why something was built a certain way, audit decision quality, or learn from past reasoning. Use when the user says "why did we," "look back," "thinkback," "decision review," "was that the right call," "what were we thinking," "review our decisions," "decision audit," "trace back," "replay thinking," or when revisiting old code and needing to understand the reasoning behind it.

When the user needs to plan a complex feature, system, or project that requires deep creative exploration before execution. Use when the user says "plan this," "let's design," "how should we build," "architect this," "brainstorm then plan," "ultraplan," "deep plan," "think this through," or when a task is too complex for a simple plan. Combines creative brainstorming with multi-agent technical validation to produce plans that are both imaginative and executable.

When contributing to public or open-source repositories where AI attribution should be hidden. Use when the user says "stealth mode," "undercover," "no AI attribution," "hide AI," "contribute anonymously," "clean commits," "no co-authored-by," or when working on repos where AI involvement should not be visible. Strips all AI fingerprints from commits, PRs, code comments, and generated content.