kiss

KISS

Keep It Simple — honestly. Look at what is used or planned (goals, processes, systems, workflows), judge whether it is overly complex, and only then recommend simplification. If complexity is earned, say so and protect it.

Simple here means:

1. A straightforward DAG from now → done (clear dependencies, few parallel critical paths, obvious stop).
2. Right-sized tasks — not micro-chores that thrash context, not epic blobs that hide risk.
3. Least mechanism that still serves the vision and constraints (safety, scale, compliance, team reality).

Operating contract

Shared by every DecisionNerd/dev-skills skill. Canonical copy: handbook/concepts/14-operating-contract.md.

• Define done first. Before using tools, write the finish line in one or two lines: acceptance criteria (existing issue AC, BDD scenarios, tests, or contract; otherwise propose them and where they should live) and the evidence that will prove them. Re-check it before reporting done. Never report done on unverified work.
• Requested scope sets the finish line. A question ("is it ready?", "why is it broken?", "what next?") ends with the answer and a Next: line naming the exact next invoke. An outcome request ("fix", "finish", "land", "#42") continues through the chain (diagnose → fix → test → check-readiness → merge-it) until the outcome or a real blocker. Do not end a turn with "Do you want me to…?" for in-scope, in-repo work.
• Stop only for real blockers. Stop and ask only when you cannot continue without the user, or before: deleting data or unmerged work, force-push or history rewrite, changing anything outside this repository (GitHub objects, deployments, live data, production or paid resources, external services), or leaving the requested scope, unless the user's request already named that exact action. Keep the harness's permission prompts for risky commands. Otherwise keep going and put status notes in the same message as the next action.
• Ask well, once. For a genuine question use the harness's structured question tool when it has one (Claude Code: AskUserQuestion; Codex: request_user_input when the current mode supports it; T3 Code: the question tool) with concrete options; otherwise plain text with numbered options. Treat the answer as settled; do not re-open earlier verdicts, plans, or answers unless asked.
• Fan out when work is parallel. Use subagents for independent reads (repo survey, evidence gathering, per-option research, per-area audits) and independent verification (a reviewer that did not write the change). Writes stay single-owner per path set and sequential. Brief every child with goal, done-when, constraints, must-not, and return shape; verify each result before consolidating. Use Claude Code's Workflow tool only when orchestrating many subagents truly needs it; it is expensive. T3 Code: its delegated subagents — with optional worktree isolation for independent verification — cover this without the Workflow tool.
• Pick the model tier per child; defer to routing config. If the harness or user config already routes subagents, follow it and do not pass a model. Otherwise pick a tier per child: mechanical search or inventory → fast/cheap; implementation and evidence gathering → mid; planning, review, adversarial verification → top. Tier names differ per harness (Claude Code: haiku / sonnet / opus or fable; T3 Code: inherit by default or pick from its live provider/model catalog). Set the choice with the harness's spawn model param or agent frontmatter model: (per-harness details: skills/agents/references/sub.md). Where the harness cannot choose, children inherit or the harness picks; say which in the status note.
• Keep a checklist on long runs. For more than about five steps or work that crosses skills, keep TASKS.md at the repo root and tick items as they finish. Do not commit it unless the repo already tracks one.
• Close every run with three headings. Blocked on me (the one genuine question or blocker, else "none"); Changed (files, commits, GitHub objects, deploys, else "nothing"); Found (evidence, verdict, and Next: <exact invoke>).

Commands

Target: path, issue/PR/milestone, doc, plan paste, agent design, or “current work.”

Routing

1. No args → audit on current git/docs scope (brief survey first).
2. First word is a command → that lens; remainder is target.
3. Clear intent (“is this architecture too much?”, “simplify the rollout plan”) → map and run.

Read-only by default. Do not delete systems, rewrite plans, or cut scope until the user accepts a recommendation.

Hard rules

1. No automatic reductionism. Complexity that buys safety, clarity, scale, or irreversible-risk control can stay. Say keep when keep is right.
2. Always present benefits and risks of simplifying (and of not simplifying).
3. Prefer one primary recommendation: simplify / keep / reshape (right-size without net removal).
4. If recommending simplify, show the simpler DAG and what was removed or merged — and what must remain.
5. Tasks in the DAG must pass the Goldilocks test: each node is finishable in one focused session (or clearly “part N”), has a done-when, and is not a fake split of a single thought.

Workflow (audit and siblings)

1. Scope the target

Name what you’re kissing: goal set, process, system, plan, or flow. Skim evidence (docs, code layout, issue graph, agent config, user’s paste). If scope is fuzzy, ask one clarifying question (structured question tool when available, else numbered options) or run a 30-second recon-style git glance. For large targets (multiple areas or packages), one subagent per lens/area (fast tier) in parallel; an independent top-tier subagent argues "keep" before you recommend "simplify"; verify each before consolidating.

2. Describe current shape

• Intent — what success is
• Pieces — goals / steps / components / roles
• Coupling — what depends on what (sketch a DAG)
• Tax — coordination cost, cognitive load, failure modes, time-to-value

3. Complexity verdict

Choose one:

Do not invent “overly complex” to have something to cut.

4. Risks & benefits

Always fill both columns:

If we simplify

• Benefits: …
• Risks: …

If we keep as-is

• Benefits: …
• Risks: …

5. Recommendation

One primary call:

• Simplify — what to remove/merge/replace; simpler DAG; migration notes
• Keep — what complexity is doing; what not to touch; optional light cleanup only
• Reshape — right-size tasks / clarify DAG / rename stages without net capability loss

Optional alternates (≤2). Hand off: refactor-it, issues narrow, milestones narrow, agents design, research-it, idk-now, DocSlime/ProductFeeling if product-shaped.

Output shape

## KISS: <target>

**Lens:** audit | goals | process | system | plan | flow

**Current shape**
- Intent: …
- Pieces: …
- DAG (sketch): …
- Tax: …

**Verdict:** overly complex | earned complexity | wrong-shaped
**Why:** <3–5 lines>

**Simplify?**
| | Benefits | Risks |
| --- | --- | --- |
| Simplify | … | … |
| Keep as-is | … | … |

**Recommend:** Simplify | Keep | Reshape
**Do next:** <concrete steps or skill invokes>
**Simpler / clearer DAG:** <mermaid or bullets, if recommending change>

**Blocked on me**
(none, or the one genuine question or missing decision)

**Changed**
(nothing — read-only; or what was reworded/removed if reshape was applied)

**Found**
Verdict, recommendation, simpler DAG if applicable. Next: `<exact invoke — e.g. refactor-it, issues narrow, agents design>`.

Goldilocks task sizing

Merge micro-tasks; split epics at real dependency or risk boundaries — not at arbitrary checklist length.

Related skills

• Lost on what matters → idk-now
• Tactical scout → recon
• Code structure only → refactor-it
• Issue/milestone scope → issues narrow / milestones narrow
• Agent stacks → agents analyze|design|optimize (slap first if thrashing)
• Docs bloat → DocSlime kiss / document-it as appropriate

Grounding

This skill’s TTPs are grounded in current engineering baselines (DORA, GitHub Docs, Fowler/Beck, Google SRE & SWE book, OpenTelemetry, OWASP LLM / NIST AI RMF, Diátaxis — see handbook sources.md).

Beck’s simple-design order (pass tests → reveal intent → remove duplication → fewer elements) and Gall’s Law: simplify when complexity is unearned, not by deleting needed safety. DORA loosely coupled architecture rewards deployable simplicity.

Handbook card: handbook/practices/kiss.md.