Stats
Actions
Tags
From bench-cowork
Use a Hammer/Anvil workflow on a complex or ambiguous task. Hammer widens the search space (exploration, architecture, critique, de-risking). Anvil narrows it (implementation, repair, verification). Use for dual-pass work, ambiguous scope, or risky changes where the cost of a wrong first pass is high.
How this skill is triggered — by the user, by Claude, or both
Slash command
/bench-cowork:hammer-anvilThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
Use this skill when a task is large, ambiguous, cross-cutting, or easy to get wrong on the first pass.
Use this skill when a task is large, ambiguous, cross-cutting, or easy to get wrong on the first pass.
Hammer widens the search space and returns a small, actionable plan.
Anvil narrows the search space and aims for the smallest safe patch.
Use Hammer/Anvil especially for:
## Surfaces affected
- <file / package / system>
## Blast radius
<what else might break>
## Risks
<concrete failure modes>
## Plan (smallest viable)
1. <step>
2. <step>
## Open questions
<what we haven't decided>
Just the diff. No narrative unless the user asks. Show:
For Firebase/App Hosting, rollout, deployment, production, Cloud Build, or CI/CD work, Anvil must first read docs/operations/app-hosting-rollout.md and keep production App Hosting rollouts exact-SHA and GitHub environment-gated through .github/workflows/apphosting-rollout.yml. Anvil must not trigger production unless the user explicitly asks for a production rollout and the approval path is clear. Do not add TURBO_TOKEN App Hosting YAML references until the Secret Manager secret exists and backend access has been granted.
npx claudepluginhub benchagi/bench-cowork --plugin bench-coworkProvides CDSS development patterns for drug interaction checking, dose validation, clinical scoring (NEWS2, qSOFA), and alert classification integrated into EMR workflows.