From workflow
Choose the right workflow pattern for a task based on Anthropic's agent research. Use before starting any non-trivial work to pick the optimal approach.
npx claudepluginhub artmin96/forge-studio --plugin workflowThis skill is limited to using the following tools:
Guides Next.js Cache Components and Partial Prerendering (PPR) with cacheComponents enabled. Implements 'use cache', cacheLife(), cacheTag(), revalidateTag(), static/dynamic optimization, and cache debugging.
Migrates code, prompts, and API calls from Claude Sonnet 4.0/4.5 or Opus 4.1 to Opus 4.5, updating model strings on Anthropic, AWS, GCP, Azure platforms.
Facilitates interactive brainstorming sessions using diverse creative techniques and ideation methods. Activates when users say 'help me brainstorm' or 'help me ideate'.
Based on Anthropic's "Building Effective Agents" research. The right pattern depends on the problem — using the same approach for everything is the #1 efficiency killer.
Analyze the task described in $ARGUMENTS and recommend ONE pattern:
When: Single file, clear scope, you can describe the diff in one sentence. Examples: Typo fix, rename a variable, add a log line, small bug fix. Action: Skip planning. Just do it. Verify with a quick test or build.
When: Multi-file change with known files. Clear what needs to happen. Examples: Add a new API endpoint, refactor a function used in 3 places. Action: Explore → Plan (list exact files + changes) → Implement → Verify.
When: Uncertain scope. You don't know which files are involved yet. Examples: "Fix the auth bug" (which auth? where?), "Improve performance" (of what?). Action: Use a subagent to explore the codebase first. Based on findings, pick pattern 1 or 2.
When: Large feature with independent subtasks. Multiple files that don't depend on each other. Examples: Add validation to 10 API endpoints, migrate 5 components to new pattern. Action: Create plan, break into independent tasks, dispatch subagents in parallel.
When: Quality-sensitive work where the first version won't be good enough.
Examples: Complex algorithm, security-critical code, public API design.
Action: Implement → Self-critique → Iterate → Verify. Use /challenge from self-critic plugin.
After choosing a workflow pattern, recommend a retrieval strategy — what information to gather and how. Different task types benefit from different retrieval approaches (Meta-Harness, arXiv 2603.28052: domain-specific routing outperforms one-size-fits-all retrieval).
git log --oneline --since="2 weeks" -- <file> — recent changes that may have introduced the buggit blame <file> on the broken section — who changed what, when~/.claude/traces/*.jsonl)grep -rn "functionName\|ClassName" --include="*.php"git log --oneline --diff-filter=M -20 -- <file> — understand change velocitygit log --all --oneline --grep="perf\|slow\|optimize"RECOMMENDED PATTERN: [Pattern Name]
REASON: [One sentence why this pattern fits]
RETRIEVAL: [Which retrieval strategy above + specific commands to run]
FIRST STEP: [What to do right now]
Key principle from Anthropic: "Start simple. Add complexity only when it demonstrably improves outcomes."