working-backwards-critic

Stub for an independent Working Backwards critic in Cursor. Products own the lock and hook. Use when reviewing a press release, FAQ, or requirements package.

SKILL.md

GitHub

Working Backwards Critic

This is an adapter stub. Product repos may replace it with a local skill and hook. Do not overwrite a drifted product copy on sync.

Canonical sources (do not fork into apps/):

  • templates/working-backwards/critic-rubric.json
  • templates/working-backwards/agent-role-contracts.md

Rules

  • Independence is a new session, not a different model. The drafting chat must not issue PASS.
  • The critic writes only verdict files (critic-review.md, session state). It does not edit writer artifacts.
  • Hooks may guard or remind. Hooks must never write PASS.
  • Products keep their own lock-file and edit-guard hook. See apt-commerce .cursor/skills/working-backwards-critic/ as a reference implementation.

Procedure

  1. Load the versioned rubric and include the version in the verdict.
  2. Read the package in stage order.
  3. Score required dimensions. Unwritten later stages stay PENDING.
  4. Distinguish [OPEN] from [BLOCKER].
  5. Return verdict, failing dimensions, evidence, and suggested fixes. Do not revise the artifacts.

Related in other

limits-colimits

Context management for Claude Code. Hooks maintain state via ledgers and handoffs. MCP execution without context polluti...

by parcadei
3,392

ddttom-allaboutv2-claude-skills-block-collection-and-party

The Block Collection and Block Party are repositories for existing AEM blocks, build tools, code snippets, integration p...

by mattnigh
20

nodnarbnitram-claude-code-extensions-claude-skills-esphome-confi

This skill should be used when the user asks to "create an esphome config", "set up an esp32 device", "configure esphome...

by mattnigh
20