dotfiles

Use when a request needs product/design exploration, unclear requirements, multiple viable approaches, UI/UX choices, or new behavior whose intent is not yet bounded.

Use when 2+ independent tasks, failures, research lanes, exploration lanes, or path-scoped investigations can run without shared state, shared files, or sequential dependency.

Use when executing a written implementation plan inline in the current session with checkpoints and verification.

Use when adding or changing a capability surface such as a public function, API, command, prompt, workflow, schema/helper, policy, or reusable instruction.

Use when implementation is complete, verification passes, and the remaining decision is how to integrate, merge, PR, clean up, or finish the branch.

Use when creating or revising docs, plans, records, PR text, handoffs, or other prose that should be sparse, direct, and low-context.

Use when changing prompt templates, AGENTS.md, CLAUDE.md, MCP/tool instructions, skills, governance docs, or any instruction surface where verbosity, duplicated doctrine, hidden exemptions, unclear provenance, or weak enforcement could affect agent behavior.

Use when making or evaluating any behavior, plan, prompt, skill, code, test, or operational change that needs purpose, proof, consumer, or falsifier clarity.

Use when changing production behavior, tests, policy gates, validators, docs, or exceptions where correctness, hidden debt, magic constants, stale contracts, or silent failures could ship.

Use when receiving code review feedback, before implementing suggestions, especially if feedback seems unclear or technically questionable - requires technical rigor and verification, not performative agreement or blind implementation

Internal guidance for presenting redteam panel JSON back to the user

Cross-model adversarial review + decision review. You orchestrate: translate user intent to panel flags, resolve doc names, pick lineages from models.json. Read-only JSON verdicts from scripts/panel.mjs + scripts/runner.mjs.

Use when completing tasks, implementing major features, or before merging to verify work meets requirements

Use when the operator says diag, diag sf, run diag, do a diag, check sf, scan sf, sf scan, sf check, scan traces, scan logs, scan spans, scan journal, anything wrong, anything flapping, is sf ok, is sf healthy, debug sf, troubleshoot sf, sf debug, sf health, what's happening, status check, quick check, eye on sf, debug scan, or asks to investigate any SF runtime anomaly, error, stall, or flapping in .sf/, the sf-server pod, Laminar spans, or cluster logs before changing code.

Use when executing a written implementation plan, or dispatching 2+ independent implementation tasks (research, audit, exploration, or coding lanes), through implementer/review/fix/integration gates.

Use when encountering a bug, test failure, production incident, unexpected behavior, performance issue, build failure, or integration failure before proposing fixes.

Use when implementing any feature or bugfix, before writing implementation code

Use when feature work needs workspace isolation, branch-safe edits, or a written plan should not run in the current checkout.

Use when starting any conversation or task, before clarifying, inspecting files, planning, editing, or answering.

Use when about to claim work is complete, fixed, passing, ready to commit, or ready for PR, especially after code, docs, config, or validator changes.

Use when you have a spec or requirements for a multi-step task, before touching code.

Use when creating, editing, pruning, or verifying skills before deployment.