Skip to content
FunCoding

Search

Search docs, Skills and MCP

deps-update

Bump outdated Hex deps — inventory, snapshot changelogs, update, fix breaks, split reviewable PRs (patches bundled, majors solo). Use to upgrade/bump Elixir dependencies or when versions fall behind. NOT for deps.get failures (/phx:investigate).

AI 与智能体565plugins/elixir-phoenix/skills/deps-update/SKILL.md

Install

Send this to Claude Code, Codex or Cursor. The agent checks the Skill for safety first and installs it only after you confirm.

读取 https://funcoding.ai/skills/oliver-kriska/claude-elixir-phoenix/deps-update/install.md ,按里面的步骤帮我安装这个 Skill。

SKILL.md

Dependency Update (Freshness)

Inventory → update → fix breaks → grouped PRs. This is the only MUTATING deps skill: it edits mix.exs, mix.lock, and source. Security scanning stays in /phx:deps-audit; the vet ledger stays in /phx:deps-vet.

Usage

/phx:deps-update                       # inventory + interactive scope pick
/phx:deps-update --scope patch         # bundle all patch bumps, one PR
/phx:deps-update --pkg phoenix_live_view   # one package (+ coupled group)
/phx:deps-update --dry-run             # inventory only, no changes

Iron Laws

  1. NEVER cross a major version without an explicit mix.exs edit — mix deps.update stays within requirements. Edit the constraint first; add override: true only when mix hex.outdated <pkg> shows a transitive consumer blocking. One major per PR
  2. ALWAYS snapshot the changelog delta BEFORE updating — capture deps/<pkg>/CHANGELOG.md, then delta via mix hex.package diff. Never update blind
  3. NEVER claim an update is safe without verification — run /phx:verify (compile --warnings-as-errors + test). "Compiles" ≠ "works"
  4. ALWAYS move coupled packages together — Phoenix core, Ecto, Ash, Oban, telemetry families update in the SAME step/commit (see ${CLAUDE_SKILL_DIR}/references/coupled-groups.md)
  5. NEVER commit a partial bump — mix.lock + mix.exs edits + (for Phoenix-family) assets/package-lock.json in ONE commit
  6. HAND OFF security to /phx:deps-audit — run it on the lock diff before any PR; don't reimplement audit rules
  7. hex.outdated exit 1 is normal — it means "deps are outdated", not failure. Capture with || true

Workflow

Phase 0: Discover

Read mix.exs: deps list, umbrella (apps_path:), git/path deps, private orgs (organization:/repo: in tuples), Phoenix/Ash presence. Create scratch dir .claude/deps-update/{YYYY-MM-DD}/.

Phase 1: Inventory

mix hex.outdated --all || true — parse the text table (no JSON exists; see ${CLAUDE_SKILL_DIR}/references/update-mechanics.md). Classify each row patch/minor/major by semver delta; Update not possible = blocked major (mix.exs constraint). Write inventory.md to scratch. Render grouped table: Patch / Minor / Major / Blocked / Git-deps (manual). --dry-run stops here.

Phase 2: Scope (AskUserQuestion)

Present groups with counts and risk. Default recommendation: "Patches (N) — low risk, bundle into one PR". --scope/--pkg flags skip the prompt. When ≥2 members of a coupled group are outdated, force them into one step even under a narrower scope.

Phase 3: Per-Package Update Loop

For each selected package, in coupled-group order:

  1. Snapshot deps/<pkg>/CHANGELOG.md → scratch/before/
  2. Update — patch/minor: mix deps.update <pkg> [coupled...]; major: edit mix.exs constraint (+ override: true if needed), then mix deps.update <pkg>
  3. git diff mix.lock → the REAL {pkg, old, new} set (hex.outdated says what could change; the lock diff says what did)
  4. Changelog delta: mix hex.package diff <pkg> <old>..<new> — keep the CHANGELOG hunk. Empty → gh api repos/{o}/{r}/releases fallback → compare-URL note (see ${CLAUDE_SKILL_DIR}/references/changelog-sources.md)
  5. Write scratch/{pkg}-{old}-{new}.md
  6. Phoenix-family in the diff + assets/package.json exists → npm install --prefix assets, stage assets/package-lock.json with the same commit

Phase 4: Verify

Run /phx:verify. On failure → Phase 5; else Phase 6.

Phase 5: Breaking-Change Fixes

Read the changelog deltas for "breaking"/"removed"/"deprecated" + the compile/test errors. Fix source (apply the sibling-file check). Re-verify.

Phase 6: Security Handoff

Run /phx:deps-audit on the working mix.lock diff (its Mode B default). BLOCK findings → surface and offer /phx:deps-vet <pkg> <ver> for accepted risks. Never skip this before a PR.

Phase 7: Group, Commit, PR

Apply the splitting strategy (${CLAUDE_SKILL_DIR}/references/pr-strategy.md): patches bundled, minors by area, majors solo, coupled groups always together. PR bodies cite the changelog excerpt, the https://diff.hex.pm/diff/<pkg>/<old>..<new> link, verification result, and the deps-audit risk band. Stage lock + mix.exs + package-lock together.

Integration

/phx:deps-update (mutating) → /phx:deps-audit (security, Mode B)
        │                              │ BLOCK → /phx:deps-vet (ledger)
        └→ /phx:verify (gate) → grouped commits / PRs

References

  • ${CLAUDE_SKILL_DIR}/references/update-mechanics.md — hex.outdated parsing, update vs unlock+get, majors, lock-diff
  • ${CLAUDE_SKILL_DIR}/references/changelog-sources.md — hex.package diff, gh fallbacks, private orgs
  • ${CLAUDE_SKILL_DIR}/references/coupled-groups.md — must-move-together groups + edge cases
  • ${CLAUDE_SKILL_DIR}/references/pr-strategy.md — grouping rules, area buckets, PR template, scratch layout

Similar Skills

brand-guidelines
anthropics/skills180k

brand-guidelines

Applies Anthropic's official brand colors and typography to any sort of artifact that may benefit from having Anthropic's look-and-feel. Use it when brand colors or style guidelines, visual formatting, or company design standards apply.

AI & agents

internal-comms
anthropics/skills180k

internal-comms

A set of resources to help me write all kinds of internal communications, using the formats that my company likes to use. Claude should use this skill whenever asked to write some sort of internal communications (status reports, leadership updates, 3P updates, company newsletters, FAQs, incident reports, project updates, etc.).

AI & agents

template-skill
anthropics/skills180k

template-skill

Replace with description of the skill and when Claude should use it.

AI & agents

mcp-builder
anthropics/skills180k

mcp-builder

Guide for creating high-quality MCP (Model Context Protocol) servers that enable LLMs to interact with external services through well-designed tools. Use when building MCP servers to integrate external APIs or services, whether in Python (FastMCP) or Node/TypeScript (MCP SDK).

AI & agents

algorithmic-art
anthropics/skills180k

algorithmic-art

Creating algorithmic art using p5.js with seeded randomness and interactive parameter exploration. Use this when users request creating art using code, generative art, algorithmic art, flow fields, or particle systems. Create original algorithmic art rather than copying existing artists' work to avoid copyright violations.

AI & agents

academy-guide
anthropics/skills180k

academy-guide

Stop and check this skill before finishing any reply to a question about how to use Claude or a Claude product — it recommends matching courses, tutorials, and use cases from Claude Academy (academy.claude.com), Anthropic's learning hub. Trigger on: "how do I", "how can I", "getting started with", "what can Claude do", "teach me", "learn to use"; questions about artifacts, projects, skills, plugins, connectors, MCP; requests about rolling Claude out to a team, class, or organization; and any ask for training materials, onboarding content, or learning resources. Use it when the user is learning how to use a feature or product — not when they are mid-task and just want the task done. This skill composes with other skills: after consulting product documentation to answer how a Claude feature works, also check here for a matching course or tutorial — a docs-grounded answer and an Academy recommendation belong together. Only recommend on a strong match; never invent Academy content.

AI & agents