Pricing
The free plan catches obvious risk in every diff — forever. Pro adds an AI layer that checks whether the agent actually built what you asked.
Catch obvious risks before every commit
Unlimited readiness checks
Run as many checks as you want against any local git repo. No limits, no usage caps.
Diff risk scan
Every check scans your uncommitted diff for deleted tests, hardcoded secrets, dependency changes, placeholder code, and oversized diffs.
Test runner integration
Configure a test command and AgentReady runs it as part of the check — pass or fail is included in the verdict.
Repair prompt
Every check generates a repair prompt you can paste directly into Cursor or Claude to guide the fix.
Saved report history
All check results are saved locally. Browse past verdicts per project at any time.
AgentForge detection
AgentReady detects your project's tech stack and surfaces which AI context files are present.
Verify the agent built the right thing
Everything in Free
All free checks run as part of every Pro check — you get the full risk scan plus the AI layer on top.
AI alignment review
Claude reads your diff against your feature description and answers: did the agent actually build what you asked, or did it drift?
Unrelated file detection
Surfaces files in the diff that have nothing to do with the feature you requested — a common sign of agent hallucination.
Scope creep analysis
Identifies changes that go beyond what was asked. Useful for keeping agents on task across multi-step sessions.
Misleading UI copy detection
Catches text in the UI that doesn't match the feature you asked for — the kind of thing that slips past diff review.
Suggested fixes
Each Pro check returns specific, actionable guidance on what to correct before committing.
AgentNarrator
Generate a commit message and PR description from the check's spec and results — copy it straight into your terminal or GitHub.
AgentForge generation
Generate AI context files for your stack — writes .cursorrules and AGENTS.md to your repo so agents have the right context from the start.
Free checks run locally on every diff. Pro adds an AI layer on top.
Deleted test coverage
Did the agent remove tests in the process of making changes?
Hardcoded secrets & tokens
Credentials, API keys, or tokens left directly in the diff.
Dependency modifications
New packages added, versions changed, or config files modified.
Placeholder & stub code
TODO comments, unimplemented stubs, or incomplete logic in production paths.
Oversized diffs
Unusually broad changes that are risky to review or hard to roll back.
Console log leakage
Debug output left in the code that shouldn't reach production.
Did the agent build the right thing?
Claude reads the full diff against your spec and returns an aligned or not-aligned verdict.
Did it touch files outside the scope?
Changed files that appear unrelated to what you asked — flagged with explanation.
Did it go beyond what was requested?
Changes that exceed the scope of the feature, including refactors you didn't ask for.
Does the UI copy match the feature?
Text in the UI that contradicts or misrepresents the feature being built.
What specifically needs fixing?
A list of targeted suggestions for each alignment issue found in the diff.
Commit message & PR description
AgentNarrator generates these from the check's spec and results.
Download the Mac app, open any git repo, and run your first check in under a minute.
Free forever. No credit card required.