Security Testing Policy

Before launching any mission, you must define scope precisely and hold documentary proof that the assets are in scope. Scope should include targets, exclusions, approved techniques, timing constraints, production restrictions and any contractual conditions that apply.

• Maintain clear written authorisation for each third-party asset owner.
• Refresh authorisation when scope, timing or methods materially change.
• Stop immediately if authorisation is revoked, expires or becomes uncertain.

You are responsible for configuring and enforcing the rules of engagement for each mission. Sentinelle may assist with guard-rails, but it does not replace your operational judgment or legal review.

• Define acceptable techniques before execution begins.
• Document prohibited systems, ports, paths, credentials and business processes.
• Use rate limits and maintenance windows suited to the target environment.
• Escalate internally before moving from discovery to intrusive validation where appropriate.

Security testing can create traffic spikes, trigger defensive systems, alter state, lock accounts, interrupt services and otherwise impact availability or integrity. These outcomes may occur even when the engagement is lawful and carefully scoped.

Sentinelle does not guarantee non-disruption. You are solely responsible for deciding whether to test production systems, when to test them, and what safeguards must be in place. You assume all risk arising from execution of missions you authorise.

• Prefer lower-risk environments first when that matches the objective.
• Use production only where justified by the engagement and tolerated by the asset owner.
• Coordinate rollback, alerting and stakeholder notification plans before high-impact validation.

Capture only the evidence reasonably necessary to prove the authorised issue. Avoid collecting unrelated personal data, secrets, customer content or other sensitive material unless doing so is necessary and authorised for the engagement.

• Redact or minimise sensitive data in reports where possible.
• Restrict access to findings and proof-of-impact to personnel with a need to know.
• Follow your internal disclosure, retention and escalation policies for critical findings.

You must stop or pause a mission immediately if you observe unexpected disruption, legal uncertainty, scope drift, customer harm, third-party complaints, or evidence that the mission may exceed the intended authorisation.

• Stop when defensive controls signal distress or instability beyond expected tolerances.
• Escalate immediately when critical vulnerabilities affect public safety, shared infrastructure or regulated data.
• Do not resume until the operational and legal basis is re-confirmed.

Sentinelle may apply default throttles, usage reviews, mission gating, output controls and other safeguards to reduce misuse and collateral harm. These controls are not guarantees and may be changed or tightened over time.

Where we reasonably believe a mission or account creates legal, security or operational risk, we may pause execution, revoke access or require additional proof of authorisation before permitting further testing.

Questions about safe operation, scope controls or testing expectations can be sent to security@sentinelle.ai or legal@sentinelle.ai.