Brief and Plan Complex Tasks

This skill runs a pre-task intelligence briefing that activates all relevant agents in the ecosystem before executing a requested task, enriching context with parallel multi-agent analysis, a realistic time estimate, a map of likely and unlikely problems, and an execution plan with contingencies - on the premise that skipping this briefing is like performing surgery without a pre-operative exam, and 30-60 seconds of parallel analysis eliminates hours of rework. Phase 1 classifies the task into Simple (answer a question, small edit - minimal briefing), Moderate (create a file, modify a skill, install a dependency - normal scan+match+estimate), Complex (new skill creation, API integration, architecture, refactoring - the full protocol), or Critical (irreversible actions, deploys, deletes, infra changes - maximum briefing plus explicit confirmation), skipping the full briefing entirely for Simple tasks. Phase 2 runs a parallel registry scan and skill-matching script against the task description, escalating to full orchestration if two or more skills match. Phase 3 briefs 3-5 of the most relevant specialized agents with targeted questions (a security agent on attack vectors/exposed data/irreversible actions, a quality agent on redundant skills or pattern compliance, an orchestration agent on what already solves part of the task, a complexity agent on computational cost, a context-continuity agent on relevant prior-session context, a legal agent on regulatory implications) - deliberately not consulting every agent blindly. Phase 4 builds an honest, step-by-step time estimate with a 20-30% contingency buffer, distinguishing agent execution time from user wait time, never underestimating to please the user, and explaining low confidence when it occurs. Phase 5 maps problems across three tiers: Probable (80%+ likely, resolved proactively before starting - e.g. validating SKILL.md YAML before install, checking API auth before calling, running git status before touching version control), Possible (30-70% likely, monitored during execution with a defined alert signal and response), and Critical-but-unlikely (under 10% probability but high impact like data loss or credential exposure, requiring a preventive backup and rollback plan). Phase 6 synthesizes everything into a final pre-execution briefing: collected agent insights, a step-by-step execution plan with per-step time and rationale, total time and confidence level, pre-resolved problems, verification checkpoints after each step, and a rollback plan for critical steps. It explicitly complements rather than replaces a separate agent-orchestrator skill - the orchestrator identifies which skills to use (routing), while this skill enriches how/when/with-what-risk to use them (briefing); both should activate together. A full worked example demonstrates the complete output for "create a Stripe integration skill," including specific security findings (API keys must never enter SKILL.md or git, webhooks need HMAC-SHA256 validation), a 7-step time-boxed plan totaling ~28 minutes at high confidence, and a concrete rollback plan.

Use it for pre-task briefing, planning task execution, analyzing before executing, or consulting agents in parallel before a Moderate/Complex/Critical task. Do NOT run the full briefing for one-line quick questions, pure read-only tasks (read/grep/glob with no side effects), simple iterations within an already-planned task, or when the user explicitly asks for a fast direct answer - the goal is intelligence in service of real speed, not bureaucracy for its own sake.

Inputs: the user's requested task description.
Outputs: a pre-execution briefing containing classified task complexity, consulted-agent insights, a time-boxed execution plan with confidence level, a three-tier problem map with pre-applied preventive fixes, verification checkpoints, and a rollback plan.

Agent orchestration systems and complex-task workflows that need a structured pre-execution risk/time briefing - consulting the right specialized agents in parallel - before committing to moderate, complex, or irreversible/critical actions.