Deterministic execution risk analyzer

Stress-test a project plan before Codex executes it.

Paste workflow text, reveal category evidence, then generate safer execution passes and a revised prompt.

Current classification
33
Low risk

Input editor

Paste a Codex/GPT project prompt or execution plan.

Risk score

33

Low risk

Weighted total: Scope 25%, Sequence 20%, Skill Conflict 15%, Test/QA 15%, Deployment 10%, Ambiguity 15%.

Weighted evidence

Scope & Complexity contributed 4 weighted points: Scope appears bounded for one planned execution.

Weighted evidence

Workflow Sequence contributed 7 weighted points: No rollback, branch, commit, or checkpoint language detected.

Weighted evidence

Skill Conflict contributed 2 weighted points: Skill set appears small enough for one pass.

Weighted evidence

Test/QA Readiness contributed 11 weighted points: No explicit test plan detected.

Weighted evidence

Deployment Readiness contributed 9 weighted points: Deploy request detected without production build verification.

Weighted evidence

Ambiguity/Context Cost contributed 2 weighted points: Ambiguity and context cost appear low.

Scope & Complexity

25% weighting

15

Evidence

  • Scope appears bounded for one planned execution.

Rule triggers

no-major-risk-detected

Workflow Sequence

20% weighting

35

Evidence

  • No rollback, branch, commit, or checkpoint language detected.

Rule triggers

missing-rollback-checkpoint

Skill Conflict

15% weighting

10

Evidence

  • Skill set appears small enough for one pass.

Rule triggers

no-major-risk-detected

Test/QA Readiness

15% weighting

70

Evidence

  • No explicit test plan detected.

Rule triggers

missing-test-plan

Deployment Readiness

10% weighting

90

Evidence

  • Deploy request detected without production build verification.
  • Deploy request detected without browser QA evidence.

Rule triggers

deploy-without-build-verificationdeploy-without-browser-qa

Ambiguity/Context Cost

15% weighting

10

Evidence

  • Ambiguity and context cost appear low.

Rule triggers

no-major-risk-detected

Failure points

  • Workflow Sequence: No rollback, branch, commit, or checkpoint language detected.
  • Test/QA Readiness: No explicit test plan detected.
  • Deployment Readiness: Deploy request detected without production build verification.
  • Deployment Readiness: Deploy request detected without browser QA evidence.

Recommended execution passes

  1. Pass 1: Confirm scope, non-goals, data boundaries, and deterministic scoring rules before editing UI.
  2. Pass 2: Implement and test the pure parsing and scoring engine with visible evidence for every score.
  3. Pass 3: Build the dashboard UI around verified engine output, including empty states and accessible controls.
  4. Pass 4: Run lint, typecheck, tests, and production build before browser QA.
  5. Pass 5: Perform desktop and mobile browser QA, then commit and deploy only after checks pass.

Execution checklist

  • Add production build verification before deployment.
  • Define exact deliverables and non-goals before implementation.
  • Write tests for deterministic parser and scoring rules before engine code.
  • Show evidence and rule triggers for every category and overall score.
  • Keep scoring client-side and framework-independent.
  • Run pnpm lint, pnpm typecheck, pnpm test, and pnpm build.
  • Run browser QA on desktop and mobile after the production build passes.
  • Commit a checkpoint before deployment.

Codex-ready revised prompt

Codex-ready revised prompt:

Build this in deterministic passes. Do not add AI chat, external APIs, server-side scoring, login, cloud sync, API routes, or automatic execution.

Pass 1: Create or update governance docs, confirm strict non-goals, and define pure TypeScript scoring contracts.
Pass 2: Write failing tests for parser signals, weighted scoring, evidence output, execution passes, checklist, and revised prompt generation.
Pass 3: Implement the client-side deterministic engine. Every category score and the overall score must display evidence and rule triggers.
Pass 4: Build the UI only after engine tests pass. Keep controls keyboard-friendly, accessible, responsive, and visually consistent.
Pass 5: Run pnpm lint, pnpm typecheck, pnpm test, and pnpm build. Run browser QA only after build passes.

Known failure points to guard against:
- Workflow Sequence: No rollback, branch, commit, or checkpoint language detected.
- Test/QA Readiness: No explicit test plan detected.
- Deployment Readiness: Deploy request detected without production build verification.
- Deployment Readiness: Deploy request detected without browser QA evidence.

Original request:
Build a new local-first dashboard, polish UI first, use multiple skills, deploy to Vercel, and check in browser. Include tests later.

Recent analyses

No saved analyses yet. Run an analysis to keep a local checkpoint on this browser.