Not another coding agent
Your coding agent creates changes. Sentaryn evaluates their authority.
AUTHORITY · EVIDENCE · CONTROL
SENTARYN FREE · $0 · LIVE ON GITHUB
For AI-heavy GitHub teams
Control what AI changesin your software.
Sentaryn is the independent authority for AI-generated software changes. SENTARYN Free checks whether AI-generated GitHub changes stayed inside the authority of the task — without enforcing its decisions.
Use Sentaryn free while we develop the product with early users. No credit card.
Catch when an AI coding agent changes files it was never asked to touch.
CodexClaude CodeCursorCopilotAgent neutral
RequestedWhat did the task ask?
AuthorizedWhat may the agent change?
ActualWhat actually changed?
EvidenceWhat supports the evaluation?
DecisionWhat did Shadow find?
RequestedFix authentication timeout
Authorized scope
✓ src/auth/**✓ tests/auth/**
Actual change
✓ src/auth/session.py✓ tests/auth/test_session.py✕ .github/workflows/deploy.yml
TestsPASS
Authorized scopeFAIL
Available evidence3 / 4
Shadow decisionScope expansion
Unauthorized scope expansion detected. Passing tests is not the same as staying authorized. Shadow reports the authority finding without enforcing the decision; the GitHub check conclusion is neutral.
SENTARYN Free · $0 · Available now
Observe changes. Review the decision.
Sentaryn Free is live through the Sentaryn GitHub App and GitHub workflow. Install it to see Shadow decisions for AI-generated changes in your repository.
No credit card. The current product operates in Shadow mode and does not enforce decisions.
NOW · SENTARYN FREE
- GitHub App / GitHub workflow
- Shadow evaluation of changes against task authority and scope
- Evidence-aware decisions using available evidence
- Informational checks; no decision enforcement
Independent authority · Core model
Agents create changes. Sentaryn checks their authority.
An AI-generated pull request can be functionally correct and still exceed its authorized scope. Sentaryn Free compares task authority with the actual change and considers available evidence to produce a decision you can review.
Requested→Authorized→Actual→Evidence→Decision
01
Requested scope
What did the task ask the agent to change?
02
Authorized scope
What was the agent allowed to change? Requested scope and authorized scope are distinct.
03
Actual change
What files did the agent actually change, and did those changes stay within its authority?
04
Available evidence
What evidence supports the evaluation, and what is missing?
05
Shadow decision
Review the evaluation in GitHub. Shadow reports the finding without enforcing the decision.
What Sentaryn Free evaluates
RequestedThe scope of the task
AuthorizedThe scope the agent may change
ActualThe observed repository change
EvidenceAvailable support and missing evidence
DecisionShadow evaluation; no decision enforcement
Different question, different control
Was this change authorized to happen?
Passing tests can show that code works. They do not prove that an AI agent was authorized to make every change in the diff. Sentaryn complements review, tests, and scanners by checking task authority, actual scope, and available evidence. It is not an LLM judging another LLM.
Not generic code review
Code review assesses the diff. Sentaryn asks whether the task authorized its scope.
Not SAST
Static security scanners look for security issues. Sentaryn compares the change with task authority.
Tests are evidence
A passing test supports behavior under that check. It does not establish authority for every changed file.
Why Sentaryn · Independent evaluation
Evaluate the change against the task.
Keep the same authority model across Codex, Claude Code, Cursor, Copilot, internal agents, and whatever comes next.
Scope comparison
Compare authorized scope with the actual Git change, including files outside the task.
Available evidence
Consider the evidence available for the change and identify gaps in that support.
Reviewable decisions
See the informational Shadow evaluation in your GitHub workflow.
Agent neutral
The authority question stays the same regardless of which coding agent created the diff.
Sentaryn Free · Live on GitHub
OBSERVE & EVALUATE
Authority questionDid the actual change stay within the task’s authorized scope?
Current modeObservation only; no decision enforcement
Where to reviewYour GitHub workflow
How it works · Current GitHub flow
Start with Sentaryn Free.
Install Sentaryn and review its evaluation alongside your existing GitHub workflow.
- Install the Sentaryn GitHub App.
- Select the repository.
- Keep working with your AI coding agent.
- Sentaryn evaluates the change against task authority.
- Review the Shadow decision in the GitHub workflow.
Sentaryn Free observes and evaluates in Shadow mode. It does not enforce decisions or configure merge protection.
Security & Trust
Understand the boundaries of Shadow evaluation.
Sentaryn Free provides an authority evaluation in GitHub. It does not enforce decisions or guarantee that a change is secure.
Repository selection
Select the repository for Sentaryn Free through the GitHub App installation flow.
Informational checks
Sentaryn Free publishes informational GitHub checks and does not configure merge protection. If your repository requires the Sentaryn check, a missing result can still prevent merging.
Evidence-aware evaluation
Review the decision alongside the available evidence and any gaps.
Keep your existing checks
Continue using code review, tests, and security scanners alongside Shadow evaluation.
Sentaryn Free · $0 · Live on GitHub
See what your AI agent changed.
Install Sentaryn to compare AI-generated changes with task authority and review the Shadow decision in GitHub. Sentaryn Free does not enforce its decisions.
WHAT YOU CAN USE TODAY
- Install the live Sentaryn GitHub App
- Keep using your AI coding agent
- Review scope, evidence, and the Shadow decision
- No decision enforcement
Sentaryn Free is live
Review security boundaries →
Observe and evaluate AI-generated GitHub changes. Sentaryn Free does not enforce decisions or configure merge protection.