design
The `/design` skill validates whether a proposed product goal aligns with strategic direction by analyzing it against a PRODUCT.md file across five dimensions: gap alignment, persona fit, competitive differentiation, precedent, and scope fit. Use it before discovery begins or when pressure-testing user value to ensure planned work supports the product's core mission and identified roadmap gaps rather than pursuing tangential features.
git clone --depth 1 https://github.com/boshu2/agentops /tmp/design && cp -r /tmp/design/images/gemini/skills/design ~/.claude/skills/designSKILL.md
# /design -- Product Validation Gate > **Quick Ref:** Validates that a proposed goal aligns with the product's strategic direction before discovery begins major work. Council-gated with `--preset=product`. --- ## Quick Start ```bash /design "add caching layer for CLI" # validate goal against PRODUCT.md /design --quick "add caching layer" # inline check, no council spawning /design --strict "redesign hook system" # higher threshold (avg >= 2.5) /design # infers goal from recent context ``` --- ## Execution Steps ### Step 0: Check for PRODUCT.md Locate `PRODUCT.md` in the repo root. ```bash ls PRODUCT.md 2>/dev/null ``` **If absent:** Output a warning and return PASS with note: "No PRODUCT.md found -- skipping product validation gate. Run `/product` to generate one." **If present:** Continue to Step 1. ### Step 1: Load Product Context Read `PRODUCT.md` and extract: - **Mission** -- the product's core purpose - **Personas** -- defined user types and their needs - **Gaps** -- known product gaps or roadmap items - **Competitive landscape** -- how the product differentiates If any section is missing, note it as unavailable and score that dimension conservatively (1). ### Step 2: Score Alignment Matrix Evaluate the proposed goal against five dimensions. Use the scoring rubric in [references/alignment-matrix.md](references/alignment-matrix.md). | Dimension | Score (0-3) | Rationale | |-----------|-------------|-----------| | Gap Alignment | | Does this goal address a known product gap? | | Persona Fit | | Does this serve defined personas? | | Competitive Diff | | Does this strengthen competitive position? | | Precedent | | Has similar work been done before? What can we learn? | | Scope Fit | | Is this appropriately scoped for the current phase? | Compute the average score across all five dimensions. ### Step 3: Run Council Invoke council with the product preset. See [references/product-council-preset.md](references/product-council-preset.md) for judge configuration. ``` Skill(skill="council", args="--preset=product validate design alignment for: <goal>") ``` Pass the alignment matrix from Step 2 as context to the council judges. If `--quick` flag is set, skip council spawning and perform inline assessment instead. ### Step 4: Write Design Artifact Write the design artifact to `.agents/design/`: ```bash mkdir -p .agents/design ``` Filename: `<date>-design-<goal-slug>.md` (e.g., `2026-03-30-design-add-caching-layer.md`) Artifact contents: - Goal statement - Alignment matrix with scores and rationale - Council verdict (or inline assessment if `--quick`) - Final verdict and recommendation ### Step 5: Output Verdict Determine the final verdict based on scores: | Condition | Verdict | |-----------|---------| | Average score >= 2.0 AND no dimension at 0 | **PASS** -- goal aligns with product direction | | Average score >= 1.5 OR one dimension at 0 with others strong | **WARN** -- goal has alignment concerns, review before proceeding | | Average score < 1.5 OR multiple dimensions at 0 | **FAIL** -- goal does not align with product direction | When `--strict` is set, raise the PASS threshold to average >= 2.5. Output the verdict in this format: ``` DESIGN VERDICT: <PASS|WARN|FAIL> Gap Alignment: <score>/3 -- <one-line rationale> Persona Fit: <score>/3 -- <one-line rationale> Competitive Diff: <score>/3 -- <one-line rationale> Precedent: <score>/3 -- <one-line rationale> Scope Fit: <score>/3 -- <one-line rationale> Average: <avg>/3.0 Artifact: .agents/design/<filename>.md ``` --- ## Flags | Flag | Effect | |------|--------| | `--quick` | Inline check without council spawning. Faster but less thorough. | | `--strict` | Raise PASS threshold from avg >= 2.0 to avg >= 2.5. Use for high-stakes changes. | --- ## Reference Documents - [references/design.feature](references/design.feature) — Executable spec: goal-vs-PRODUCT.md council verdict before discovery, --quick inline, --strict threshold (soc-qk4b) - [references/alignment-matrix.md](references/alignment-matrix.md) -- Scoring rubric for the five alignment dimensions - [references/project-reality-check.md](references/project-reality-check.md) -- Quick reality-check prompts before committing to a project direction - [references/product-council-preset.md](references/product-council-preset.md) -- Council `--preset=product` judge configuration and verdict rules --- ## See Also - [skills/council/SKILL.md](../council/SKILL.md) -- Multi-model consensus council - [skills/product/SKILL.md](../product/SKILL.md) -- Generate PRODUCT.md - [skills/discovery/SKILL.md](../discovery/SKILL.md) -- Discovery phase orchestrator - [skills/pre-mortem/SKILL.md](../pre-mortem/SKILL.md) -- Plan validation gate
Use Agent Mail from Codex for file leases, notifications, inboxes, and conflict prevention.
>-
>-
Use when converting markdown plans into br beads with dependencies for implementation or swarm execution.
Use when switching AI coding CLI accounts quickly to recover from subscription rate limits or OAuth friction.
>-
Use when starting non-trivial work, mining lessons, or preventing repeated mistakes with cm procedural memory.
Mine past agent sessions for working prompts, decisions, and patterns. Use when "what did I ask?", "find that prompt", session archaeology, or agent history.