extract-skill
Reverse-engineer design systems, tokens, and components from live products or screenshots
Persist design documents with branch tracking, revision chains, and cross-session discovery
$ npx -y skills add nyldn/claude-octopus --skill skill-design-lineage --agent claude-codeHow it fires
How this skill gets triggered: by you, by Claude, or both.
/skill-design-lineageContext preview
The summary Claude sees to decide when to auto-load this skill.
Persist design documents with branch tracking, revision chains, and cross-session discovery
name: skill-design-lineage description: "Persist design documents with branch tracking, revision chains, and cross-session discovery" disable-model-invocation: true
> **Host: Codex CLI** — This skill was designed for Claude Code and adapted for Codex. > Cross-reference commands use installed skill names in Codex rather than `/octo:*` slash commands. > Use the active Codex shell and subagent tools. Do not claim a provider, model, or host subagent is available until the current session exposes it. > For host tool equivalents, see `skills/blocks/codex-host-adapter.md`.
When a design introduces or changes domain terms, load `skills/blocks/domain-modeling.md`. Record each definition in the existing design lineage artifact with its evidence source, checked time, and scope. A changed definition invalidates dependent decisions. Link and reopen them instead of leaving stale implementation guidance marked current.
Persist design documents from brainstorming and planning sessions with branch tracking, revision chains, and cross-session discoverability. Design docs are immutable after creation -- new revisions supersede prior versions rather than editing in place.
All design documents are stored under:
~/.claude-octopus/designs/<project-slug>/
The project slug is derived from the current git repository name or working directory basename:
SLUG=$(basename "$(git rev-parse --show-toplevel 2>/dev/null || pwd)")
DESIGNS_DIR="${HOME}/.claude-octopus/designs/${SLUG}"
mkdir -p "$DESIGNS_DIR"Design documents follow a strict naming convention:
{user}-{branch}-design-{datetime}.mdWhere:
Example:
chris-feature-auth-refactor-design-20260321-143022.md
Resolution:
BRANCH=$(git rev-parse --abbrev-ref HEAD 2>/dev/null | tr '/' '-' || echo "no-branch")
DATETIME=$(date -u +"%Y%m%d-%H%M%S")
FILENAME="${USER}-${BRANCH}-design-${DATETIME}.md"Each design document contains YAML frontmatter followed by structured sections.
branch: feature/auth-refactor user: chris created: 2026-03-21T14:30:22Z supersedes: chris-feature-auth-refactor-design-20260320-091500.md
Fields:
# Design: [Title] ## Problem Statement [What problem does this design solve? Why does it matter?] ## Constraints [Technical, timeline, resource, or organizational constraints that bound the solution space.] ## Approaches Considered [List each approach evaluated, with brief pros/cons for each.] ### Approach A: [Name] - **Pros:** ... - **Cons:** ... ### Approach B: [Name] - **Pros:** ... - **Cons:** ... ## Recommendation [Which approach is recommended and why. Include the key trade-off that drove the decision.] ## Open Questions [Unresolved questions that may affect implementation or require follow-up.]
After a brainstorm, planning, or define workflow produces design output, save it to the standard location with metadata.
SLUG=$(basename "$(git rev-parse --show-toplevel 2>/dev/null || pwd)")
DESIGNS_DIR="${HOME}/.claude-octopus/designs/${SLUG}"
mkdir -p "$DESIGNS_DIR"
BRANCH=$(git rev-parse --abbrev-ref HEAD 2>/dev/null | tr '/' '-' || echo "no-branch")
DATETIME=$(date -u +"%Y%m%d-%H%M%S")
FILENAME="${USER}-${BRANCH}-design-${DATETIME}.md"
FILEPATH="${DESIGNS_DIR}/${FILENAME}"
cat > "$FILEPATH" <<EOF
branch: $(git rev-parse --abbrev-ref HEAD 2>/dev/null || echo "no-branch")
user: ${USER}
created: $(date -u +"%Y-%m-%dT%H:%M:%SZ")
${SUPERSEDES:+supersedes: ${SUPERSEDES}}
# Design: [Title]
## Problem Statement
[Content from brainstorm/planning session]
## Constraints
[Identified constraints]
## Approaches Considered
[Evaluated approaches]
## Recommendation
[Selected approach and rationale]
## Open Questions
[Unresolved items]
EOFDesign docs are **read-only after creation**. To revise a design, create a new document with a `supersedes` reference to the prior version.
Before writing a new design, search for related prior designs using keyword matching.
SLUG=$(basename "$(git rev-parse --show-toplevel 2>/dev/null || pwd)")
DESIGNS_DIR="${HOME}/.claude-octopus/designs/${SLUG}"
# Search for designs matching keywords (case-insensitive)
KEYWORDS="auth refactor"
grep -li "$KEYWORDS" "${DESIGNS_DIR}"/*.md 2>/dev/null | head -10Constraints:
When prior designs are found, present them to the user:
Found 2 prior designs related to "auth refactor": 1. chris-feature-auth-refactor-design-20260320-091500.md (2026-03-20) 2. chris-main-design-20260315-140000.md (2026-03-15) Would you like to review any of these before creating a new design?
If a prior design exists for the same branch, the new design document includes a `supersedes` field in its frontmatter pointing to the prior filename.
BRANCH=$(git rev-parse --abbrev-ref HEAD 2>/dev/null | tr '/' '-' || echo "no-branch")
# Find the most recent design for this branch
PRIOR=$(ls -t "${DESIGNS_DIR}"/*-${BRANCH}-design-*.md 2>/dev/null | head -1)
if [[ -n "$PRIOR" ]]; then
SUPERSEDES=$(basename "$PRIOR")
fiThis creates a linked chain of revisions: `
Every AI model has blind spots. Claude Octopus supports twelve external provider integrations — Codex, Antigravity CLI, Copilot, Qwen, Ollama, Perplexity, OpenRouter, OrcaRouter, OpenCode, Cursor CLI, Grok, and Kimi Code — alongside the built-in Claude Code
Repo: nyldn/claude-octopus
Reverse-engineer design systems, tokens, and components from live products or screenshots
Multi-AI requirements scoping using available external providers (Double Diamond Define phase). Priority triggers: octo define, octo scope, co-define,…
Multi-AI validation, scoring, and review using available external providers (Double Diamond Deliver phase)
Multi-AI implementation using available external providers (Double Diamond Develop phase). DO NOT use for simple code edits, reading/reviewing code, built-in…
Multi-AI research using available external providers (Double Diamond Discover phase)
Decompose and execute large changes, migrations, or multi-issue fixes in parallel with quality gates